SANDIA REPORT

SAND2016-11399
Unlimited Release
Printed November 2016

Heat Transfer Phenomena in

Concentrating Solar Power Systems
Kenneth M. Armijo (PI) and Subhash L. Shinde (PM)

Prepared by
Sandia National Laboratories
Albuquerque, New Mexico 87185 and Livermore, California 94550

Sandia National Laboratories is a multi-mission laboratory managed and operated by Sandia Corporation,
a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's
National Nuclear Security Administration under contract DE -AC04-94AL85000.

Approved for public release; further dissemination unlimited.

Issued by Sandia National Laboratories, operated for the United States Department of Energy
by Sandia Corporation.

NOTICE: This report was prepared as an account of work sponsored by an agency of the
United States Government. Neither the United States Government, nor any agency thereof, nor
any of their employees, nor any of their contractors, subcontractors, or their employees, make
any warranty, express or implied, or assume any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process disclosed, or
represent that its use would not infringe privately owned rights. Reference herein to any specific
commercial product, process, or service by trade name, trademark, manufacturer, or otherwise,
does not necessarily constitute or imply its endorsement, recommendation, or favoring by the
United States Government, any agency thereof, or any of their contractors or subcontractors.
The views and opinions expressed herein do not necessarily state or reflect those of the United
States Government, any agency thereof, or any of their contractors.

Printed in the United States of America. This report has been reproduced directly from the best
available copy.

Available to DOE and DOE contractors from

U.S. Department of Energy
Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831

Telephone: (865) 576-8401

Facsimile: (865) 576-5728
E-Mail: reports@osti.gov
Online ordering: http://www.osti.gov/scitech

Available to the public from

U.S. Department of Commerce
National Technical Information Service
5301 Shawnee Rd
Alexandria, VA 22312

Telephone: (800) 553-6847

Facsimile: (703) 605-6900
E-Mail: orders@ntis.gov
Online order: http://www.ntis.gov/search

2
 SAND2016-11399
Unlimited Release
Printed November 2016

Heat Transfer Phenomena in Concentrating Solar

Power Systems
Kenneth M. Armijo (PI) and Subhash Shinde (PM)
Concentrating Solar Energy Technologies Department
Sandia National Laboratories
P.O. Box 5800
Albuquerque, New Mexico 87185-MS1127

Abstract

Concentrating solar power (CSP) utilizes solar thermal energy to drive a thermal power cycle for
the generation of electricity. CSP systems are facilitated as large, centralized power plants, such
as power towers and trough systems, to take advantage of economies of scale through dispatchable
thermal energy storage, which is a principle advantage over other energy generation systems.
Additionally, the combination of large solar concentration ratios with high solar conversion
efficiencies provides a strong opportunity of employment of specific power cycles such as the
Brayton gas cycle that utilizes super critical fluids such as supercritical carbon dioxide (sCO2),
compared to other solar-fossil hybrid power plants. A comprehensive thermal-fluids examination
is provided by this work of various heat transfer phenomena evident in CSP technologies. These
include sub-systems and heat transfer fundamental phenomena evident within CSP systems, which
includes receivers, heat transfer fluids (HTFs), thermal storage media and system designs,
thermodynamic power block systems/components, as well as high-temperature materials. This
work provides literature reviews, trade studies, and phenomenological comparisons of heat transfer
media (HTM) and components and systems, all for promotion of high performance and efficient
CSP systems. In addition, further investigations are also conducted that provide advanced heat
transfer modeling approaches for gas-particle receiver systems, as well as performance/efficiency
enhancement recommendations, particularly for solarized supercritical power systems.

3
 ACKNOWLEDGMENTS
Sandia National Laboratories is a multi-mission laboratory managed and operated by Sandia
Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department
of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000. This
work was primarily funded by the US Department of Energy Solar Energy Technologies Program.
This material is also based upon work partially supported by the U.S. Department of Homeland
Security under Grant Award Numbers 2012-DN-130-NF0001-0202 and HSHQDC12X00059. The
views and conclusions contained in this document are those of the authors and should not be
interpreted as representing the official policies, either expressed or implied, of the U.S. Department
of Homeland Security.

4
 TABLE OF CONTENTS

1. Introduction ................................................................................................................................ 7
2. CSP Heat Transfer .................................................................................................................... 12
2.1. Radiation Heat Transfer ................................................................................................ 12
2.2. Conduction Heat Transfer ............................................................................................. 19
2.3. Convection Heat Transfer ............................................................................................. 22
2.4. Phase Change & Latent Heat Transfer.......................................................................... 26
2.4.1. Surface Tension & Boiling .............................................................................. 27
2.4.2. Wetting Characteristics .................................................................................... 28
2.4.3. Phase Stability ................................................................................................. 28
2.4.4 Pool & Convective Boiling Heat Transfer ....................................................... 30
3.4.4.1 Pool Boiling Heat Transfer............................................................ 30
3.4.4.2 Convective Boiling Heat Transfer ................................................. 33
2.4.5 CSP Multimode Heat Transfer......................................................................... 36
3. Thermal Collectors.................................................................................................................... 38
3.1. Overview ....................................................................................................................... 38
3.2. Thermal Colelctors Analysis......................................................................................... 40
3.3. Solar Selective Surface Coatings .................................................................................. 48
3.4. Collector Performance and Efficiency .......................................................................... 50
3.5. CSP System Receiver Technologies ............................................................................. 54
3.5.1 Direct Absorption Receivers ............................................................................ 55
3.5.1.1 Cavity Receivers............................................................................ 55
3.5.1.2 Direct Absorption Particle Receviers ............................................ 61
3.5.1.3 Advanced Magnetic Falling Particle Receivers ............................ 67
3.5.1.4 Other Advanced Direct Receivers ................................................. 69
3.5.2 Indirect Absorption Receivers .......................................................................... 70
3.5.2.1 Indirect Absorption Particle Receivers.......................................... 70
3.5.2.2 Tubular Receivers.......................................................................... 75
3.5.2.3 Volumetric Receivers .................................................................... 78
4. Heat Transfer Fluids for CSP Systems ..................................................................................... 79
4.1. Overview ....................................................................................................................... 79
4.2. Synthetic Oils ................................................................................................................ 83
4.3. Molten Salts .................................................................................................................. 84
4.4. Metals ............................................................................................................................ 87
4.4.1 Potassium and NaK .......................................................................................... 89
4.4.2 Lithium ............................................................................................................. 89
4.4.3 Sodium ............................................................................................................. 90
4.4.3.1 Sodium Saftey .............................................................................. 91
4.4.3.2 Reactivity with Liquid Water ....................................................... 91
4.4.3.3 Reactivity with Air and Water Vapor ........................................... 92
4.4.4 Lead Bismuth Eutectic (LBE) .......................................................................... 93
4.5. Gases ............................................................................................................................. 94
4.5.1 Air .................................................................................................................... 94

5
 4.5.2 Helium.............................................................................................................. 95
4.5.3 Water/Steam ..................................................................................................... 96
4.5.4 CO2................................................................................................................... 97
5. Thermal Energy Storage ........................................................................................................... 99
5.1 Overview ....................................................................................................................... 99
5.2 Sensible Energy Storage ............................................................................................. 102
5.2.1 Molten Salts ................................................................................................... 103
5.3 Latent Energy Storage................................................................................................. 104
5.4 Thermochemical Energy Storage ................................................................................ 106
6. CSP Thermodynamic Power Systems .................................................................................... 109
6.1 Overview ..................................................................................................................... 109
6.2 Carnot Cycle & Thermodynamic Limitations ............................................................ 110
6.3 Rankine Cycle ............................................................................................................. 114
6.3.1 Overview and Theory .................................................................................... 114
6.4 Brayton Cycle ............................................................................................................. 116
6.4.1 Overview & Theory ....................................................................................... 116
6.4.2 Supercritical Brayton Cycles ......................................................................... 120
6.4.3 Brayton Cycle Components & Advanced Designs ........................................ 126
6.4.3.1 Reheat ......................................................................................... 126
6.4.3.2 Recuperation & Recompression ................................................. 126
6.4.3.3 Intercooling & Regeneration ...................................................... 128
6.4.3.4 Combined Cycle ......................................................................... 130
6.4.3.5 Overspray ................................................................................... 131
6.5 Stirling Cycles ............................................................................................................. 131
6.6 Other CSP Cycles ....................................................................................................... 132
7. CSP Power Block Heat Exchange Components ..................................................................... 134
7.1 Heat Exchangers ......................................................................................................... 134
7.1.1 Gasketed-Plate Heat Exchangers ....................................................................... 141
7.1.2 Shell and Tube Heat Exchangers ....................................................................... 144
7.1.3 Compact Heat Exchangers ................................................................................. 147
7.1.4 Plate Heat Exchangers ....................................................................................... 149
7.1.5 Plate-Fin Heat Exchangers ................................................................................. 150
7.1.6 Printed Circuit Heat Exchangers ........................................................................ 152
7.1.7 Spiral Heat Exchangers ...................................................................................... 153
7.1.8 Ceramic Heat Exchangers .................................................................................. 154
7.1.9 Fluidized Bed Heat Exchangers ......................................................................... 155
7.2 Recuperators ............................................................................................................... 157
7.3 Regenerators ............................................................................................................... 157
8. Concluding Remarks ............................................................................................................... 160
9. References ............................................................................................................................... 161
10. Appendix ............................................................................................................................... 186

6
1. INTRODUCTION
Solar thermally generated electricity is a low cost renewable energy source that employs highly-
absorptive collectors to gather solar radiation for facilitating temperature power cycles to produce
clean and affordable power. Overall conversion (system) efficiency of approximately 35% is
feasible with intelligent thermal management [1]. Solar thermal radiation can be collected by
different concentrating solar power (CSP) technologies for providing a high temperature heat
source. These technologies however differ depending on whether they are line or point-focusing.
There is also a distinction between either the distribution of an optical concentrator in the form of
small planar fragments and continuous curved concentrators. Contemporarily, the four principal
CSP technologies evident in commercialization are shown in Fig. 1.1 by their degree of
commercialization (left to right).

Figure 1.1. Principal CSP technologies used within industrial practice [2, 3].

In a CSP plant heat is used to operate a conventional power cycle, such as Rankine (stem engine),
Brayton (gas turbine engine) or Stirling engine [1]. Stored heat can then be used during nighttime
or overcast conditions for power generation. A simple overall schematic for a typical CSP system
can be seen in Fig. 1.1, which describes the main general elements.
Typical market categories for CSP vary between small (<100 kW), medium (<10 kW) and
large (>10 MW) where the systems can be composed of combinations of different collectors,
power cycles and also thermal storage technologies. A CSP system processes heat like
conventional power plants where the efficiency depends on the operating temperature. Therefore,
the useful energy produced will depend on solar field collection and power cycle efficiencies as
depicted in Fig. 1.2.

7
 Figure 1.2. Main components of a Concentrating Solar Power (CSP) System [4].

Concentrated solar power systems use mirrors or lenses to concentrate a large area of sunlight, or
solar thermal energy, onto a small area. Electrical power is produced when the concentrated light
is converted to heat which drives a heat engine (usually a steam turbine) connected to an electrical
power generator. Unlike the photovoltaic solar cells, converting energy from sunlight to electricity
by CSP systems is based on the application of heat engine rather than photovoltaic effect which is
directly transfer photon energy into electricity energy.

Figure 1.3. CSP system efficiency variation with operating temperature [1].

The collectors concentrate and collect solar thermal energy which is used to generate vapor that
runs heat engines to produce electricity. The theoretical maximum efficiency that can be achieved
is determined by the Carnot’s cycle, where the efficiency of a heat engine is determined by the
difference between the lowest and highest temperatures reached in one cycle. However, thermal
losses of the solar receiver increase with the fourth power of its surface temperature [5], therefore,
efficiency gains of the power block from a higher temperature can be potentially negated by the
increasing thermal losses from the receiver due to higher temperatures. Therefore, the maximum
efficiency can be increased and shifted towards higher temperatures when solar irradiation is
further concentrated by CSP collectors. However, thermal losses remain constant for the same
receiver surface area despite yielded solar power increasing. This is what allows CSP systems with
higher concentration ratios to achieve higher system efficiencies. This is illustrated by Fig. 1.4
which presents CSP systems with varying concentration ratios and their respective maximum
system efficiencies with respect to Carnot.

8
Figure 1.4. CSP technology theoretical efficiencies with respect to concentration ratio and
absorber temperature [6].

Since the 2012 SunShot Vision Study, global deployment of concentrating solar power (CSP) has
increased threefold to nearly 4,500 MW, with a similar threefold increase in operational capacity
to 1,650 MW within the United States. Fig. 5 demonstrates that this deployment has led to dramatic
cost reductions that have placed CSP well on the path to reaching the U.S. Department of Energy
(DOE) SunShot Initiative goal of 6 cents/kWh by 2020 [8]. However enhanced power conversion
systems are needed for responding to ever-increasing energy demands for the U.S. and globally.
Efficiency in CSP systems is key for producing the most power for the smallest investment. A
Brayton cycle typically operates at higher temperatures and pressures, which improves the overall
efficiency compared to more universally-used Rankine cycle systems.

Figure 1.5. Cost reductions for parabolic trough and tower technologies since the original SunShot
Vision Study (SVS) [7].

9
Brayton systems can be furthered improved through the employment of supercritical fluids such
as water, carbon dioxide and helium that can avoid fundamental boiling/condensing phase change
issues, as well as increasing Carnot temperature differences within the cycle. Incorporation of
thermal energy storage is the key advantage of CSP plants and can be used to produce electricity
when the sun is not shining. The round-trip efficiency of thermal storage can be very high (≥ 99%),
and costs are low relative to alternatives such as battery storage. SunShot goals for thermal storage
include power-cycle inlet temperature ≥ 650 °C, energetic efficiency ≥ 99%, exergetic efficiency
≥ 95%, and a cost ≤ $15/kWth.

Figure 1.6. The solar insolation (kWh/m2/day) on an optimally tilted surface. [1].

The potential for CSP implementation in any given geographic location is largely determined by
the local solar radiation potential. Fig. 1.6 shows average solar insolation distributions globally
for an optimally tilted surface during the strongest month of the year [6,7]. Regions represented
by light and dark red colors are most suitable for CSP implementation, based on irradiance,
measured in W/m2.
The daily variation of the different components depends upon meteorological and
environmental factors (e.g. cloud cover, air pollution and humidity) and the relative earth-sun
geometry. The direct normal irradiance (DNI) is synonymous with the direct beam radiation and
it is measured by tracking the sun throughout the sky. Fig. 1.3 shows an example of the global
solar radiation that is measured on a stationary two flat plate and a plate that is tracking the sun.
The measured DNI is also included and its lower value can be attributed to the fact that it does not
account for the diffuse radiation component.
In CSP applications, DNI is important in determining available solar energy as its value is
determined only by incident normal radiation, where collectors are designed to track the sun
throughout the day. Fig. 1.7 shows the daily solar insolation on an optimally tilted surface during
the worst month of the year around the world [6,7]. Regions represented by light and dark red
colors are most suitable for CSP implementation. The annual DNI value will also greatly influence
the levelized cost of electricity cost (LCOE), which will be discussed later. Typical values of DNI
at different latitudes and selected locations around the world are given in Figure 6 and Table 1.
Based on the information presented here it can be seen that desert and equatorial regions appear to
provide the best resources for CSP implementation.

10
Figure 1.7. Solar irradiance variation within a day measured on a flat plate positioned horizontal
and tracking system the sun and direct normal irradiance (DNI) [1, 8].

Equivalent radiation temperature of the sky depends on the air density and its moisture content.
When the relative humidity is high and at sea level, the sky temperature can be assumed to be the
same as ambient air temperature. However, for low relative humidity or at high altitudes, the sky
radiation temperature can be 6 to 8 ᵒC less than ambient temperature. If there is no atmosphere as
with space applications, the equivalent sky temperature approaches 0 K.
For this work, radiation and other heat transfer principles will be discussed in a
comprehensive overview for providing insight into the primary mechanisms of thermal-fluid
behavior in CSP systems. In the subsequent chapters, a broad discussion will follow that provides
the reader with greater detail of heat transfer phenomena and system-level design details for
various CSP receivers currently in operation, as well as advanced concepts and design. Following
this discussion, further heat transfer properties and general operational characteristics will be
covered for a variety of heat transfer and working fluids employed in CSP and respective power
block systems. These chapters will cover not just the ability for these fluids to transfer thermal
energy, but also to store it. This dialogue will then be followed by two chapters that provide insight
into current and advanced thermodynamic systems employed in CSP systems as well as heat
transfer characteristics of components within the power block, for enhancing performance and
reliability to further promote ubiquitous adoption of CSP for electrical energy generation.

11
2. CSP HEAT TRANSFER PRINCIPLES
Throughout the various stages of the CSP plant, heat transport is optimized to maximize
conversion of solar heat to mechanical work by reducing thermal and optical losses.
Fundamentally, an optimal CSP system design combines a relatively large, efficient optical surface
(e.g., a field of high-reflectivity mirrors), harvesting incident solar radiation and concentrating it
onto a solar receiver with a relatively small aperture area. To facilitate nominal CSP power
generation ideal solar receivers would have negligible convection and conduction losses, where
heat transfer to the rest of the solar thermal power plant is facilitated through maximum heat
exchange to a working fluid with an outlet temperature high enough to feed a heat engine that
produces electricity. For any heat transfer component, there are three principal modes of heat
transfer: conduction, convection, and radiation where any energy exchange between bodies occurs
through these modes or in combination based on temperature differences.

2.1 Radiation Heat Transfer

Radiation is energy emitted by a solid object, liquid or gas that is at a finite temperature. This
matter can also dissipate radiant energy impinging on it by reflection and can be capable of
absorption. For thermal energy processes there are several solar flux types which are illustrated in
Fig. 2.1, where it is convenient to generalize by two wavelength ranges [9] short wavelength and
long wavelength. Short wavelength radiation originates from the sun or high-temperature sources
whereas long wavelength radiation typically originates from lower-temperature sources and is
emitted in proportion to temperature to the fourth power.

a. b. c.
Figure 2.1. Radiation energy flux types which include a. direct beam radiation, b. diffuse radiation
and c. reflected radiation.

Solar radiation at normal incidence, and at a particular surface on earth is subject to variations of
extraterrestrial radiation, as well as atmospheric scattering and absorption by air molecules, water,
dust and aerosols as described by Armijo et al. [10]. An illustrative description of irradiance
attenuation from that of extraterrestrial irradiance, by various atmospheric aerosols and water, is
shown in Fig. 2.2. For PV technologies the spectral response of a particular cell material can be
susceptible to power generation reduction due to these atmospheric components; however, CSP is
much more spectrally robust as it employs a significantly broader spectrum for power generation.

12
Figure 2.2. Solar irradiance spectrum comparison between extraterrestrial and atmospherically-
attenuated irradiance with an air mass (AM) equal to unity [11].

Heat is also transferred by direct contact with solid objects (conduction), movement of air that
carries heat away from the object (convection or sensible heat), and latent heat exchange in which
heat is dissipated through a change in water from liquid to gas. The radiation that impinges on a
surface or object must be conserved by the energy radiated back to the ambient, gained by sensible
(conduction)/latent heat (convection) or thermal energy storage [458]. Accordingly, this radiation
energy balance can be defined as:

(1 − 𝑟)𝑆 ↓ +𝐿 ↓= 𝐿 ↑ +𝐻 + 𝜆𝐸 + 𝐺 (2.1)

The left side of the equation consists of absorbed solar radiation (1 − 𝑟)𝑆 ↓ and longwave radiation
(𝐿 ↓). The right side of the equation consists of the emitted longwave radiation (𝐿 ↑), sensible heat
(H), latent heat (λE) and heat exchange by conduction (G) [458]. Therefore, the net radiation
absorbed by an object is:
𝑅𝑛 = (1 − 𝑟)𝑆 ↓ +(𝐿 ↓ −𝐿 ↑) = 𝐻 + 𝜆𝐸 + 𝐺 (2.2)

Net radiation (Rn) is balanced by sensible, latent and conduction heat transfer, where all surfaces
except for those that are blackbody reflect or transmit sunlight. This amount is equal to rS, where
S is the incident radiation onto the surface where r is defined as the fraction of S that is reflected
by the surface [458]. The remainder term, (1 − 𝑟)𝑆, is the solar radiation absorbed by the surface.
Emission of longwave radiation is a second means of radiative cooling. Terrestrial objects emit
electromagnetic radiation in the infrared band at long wavelengths between 3 μm-100 μm. This
emission is proportional to temperature raised to the fourth power, where the units are with respect
to degrees Celsius:
𝐿 ↑= 𝜀𝜎(𝑇𝑠 + 273.15)4 (2.3)

where σ = 5.67x10-8 W/m2-K4 is the Stefan-Boltzmann constant and emissivity, ε of an object

generally ranges from 0.8-1.0 which is the emissivity of a blackbody. Most natural surfaces are
considered “grey” which have an emissivity less than unity. However, unlike conduction and

13
convection, which require the presence of a temperature gradient in some form of matter to
facilitate heat transfer, radiation heat transfer does not require matter and is an extremely relevant
to many industrial processes that involve combustion and solar energy. The mechanism of
radiative emission is related to the energy released as a result of oscillations or transitions of many
electrons that constitute matter. These oscillations are sustained by internal energy and therefore
the temperature of matter, where all forms of matter emit radiation [12]. For gases and
semitransparent solids, such as glass and salt crystals at elevated temperatures, emission is a
volumetric phenomenon, where the radiation that emerges from a finite volume of matter is the
integrated effect of local emission throughout the volume. For the majority of liquids and solids
radiation emitted from interior molecules is strongly absorbed by adjoining molecules, where this
emitted radiation originates from molecules that are within a distance of approximately 1 μm from
an exposed surface. Therefore, emission, or absorption, from solids or liquids into an adjoining
gas or vacuum can be assessed as a surface phenomenon [12].
Sunlight, in the broad sense, is the total spectrum of the electromagnetic radiation given
off by the Sun. Thermal radiation may be viewed as the transport of photons or the propagation of
an electromagnetic wave, where for the propagation of radiation into a medium the frequency and
wavelength are related as:
𝑐
𝜆=𝑣 (2.4)

where c is the speed of light in the medium, λ is the wavelength, and v is the frequency. For a
vacuum the propagation of radiation follows that c = 2.998x108 m/s. The complete electromagnetic
spectrum is presented by Fig 2.3, where between approximately 0.1 to 100 μm thermal radiation
is pertinent to heat transfer.

Figure 2.3. Thermal infrared radiation bands within the electromagnetic spectrum [13]

The thermal radiation band includes a portion of UV, all of the visible, as well as the IR which can
be broken into four different subsections. First, the near infrared (NIR) band extends from 0.7 to

14
1 μm, while short-wave IR (SWIR) extends from 1 to 2.5 μm, where these spectra are generally
the result of reflections from objects, which are similar radiation in the visible spectrum, however
both NIR/SWIR light are not visible to humans, and can be considered as a non-thermal form of
infrared radiation [13]. The mid-wave infrared (MWIR) region extends from 3 to 5 μm and the
long-wave infrared (LWIR) extends from 7.5 to 14 μm, where both radiation spectra bands
originate from an object’s surface thermal emissions, and not reflections. Both MWIR and LWIR
are considered the primary sources of thermal radiation. In addition, the terahertz region, which is
between 100 and 1000 μm (0.3–30 THz), is part of the very long-wave infrared (VLWIR), from
14 to 1000 μm, where this radiation can be considered as a special case of thermal infrared
radiation, sharing some properties with the infrared and others from the microwave bands [13]. In
general, emitted radiation consists of a continuous, non-uniform distribution of monochromatic
wavelength spectra, where the magnitude of the radiation at any wavelength, as well as the spectral
distribution vary with temperature and directionality of the emitting surface [12]. Although
radiation emitted by a surface propagates in all directions, radiation incident onto a surface may
come from different directions, which can influence the surface response. For a radiative emitting
surface of an arbitrary element of area dA1 a differential solid angle dω may subtend a point onto
a differential area dA2 as shown in Fig. 2.4.

Figure 2.4. Radiation emission from an arbitrary differential element dA1 into a solid angle dω
subtended by dAn at a point on dA1 [12].

This direction can be specified in terms of zenith, θ and azimuth, 𝜙 angles of a spherical coordinate
system. Additionally, if the midpoint within dA2 were a sphere, the differential solid angle dω can
be defined by a region between the rays of the sphere and measured as the ratio of the element area
dA2 on the sphere to the square of the sphere’s radius as prescribed by Eqn. 2.5.

𝑑𝐴2
𝑑𝜔 = (2.5)
𝑟2

In contrast, as shown in Fig. 2.4 incident radiation from the area dA2 can also subtend a solid angle
dω onto a surface dA1, where it can be related to a radiative flux or irradiation, which encompasses
radiation incident from all directions [12]. With respect to wavelength, spectral irradiation G

15
(W/m2-μm) is defined as the rate at which radiation of wavelength λ is incident on a surface per
unit surface area and unit wavelength interval dλ about λ, Eqn. 2.6.
2𝜋 2𝜋
𝐺𝜆 (𝜆) = ∫0 ∫0 𝐼𝜆 (𝜆, 𝜃, 𝜙)𝑐𝑜𝑠𝜃𝑑𝜔 (2.6)

where the unit solid angle dω = sinθdθdφ. If the total irradiation is taken with respect to the rate at
which radiation is incident per unit area from all directions and all wavelengths, then one is
provided:
∞
𝐺 = ∫0 𝐺𝜆 (𝜆)𝑑𝜆 (2.7)

However, if the incident radiation is diffuse, Iλ(λ) is therefore independent of θ and φ, where it
follows that:
Gλ(λ)=πIλ(λ) (2.8)

For radiation leaving a surface radiosity encompasses the reflected portion of irradiation, in
addition to direct emission as shown by Fig. 2.5. Spectral radiaosity Jλ (W/m2-μm) represents the
rate at which radiation of wavelength λ leaves a unit area of the surface, per unit wavelength
interval dλ. This parameter assumes surface radiation to leave in all directions and is related to the
intensity associated with emission e, and reflection r, 𝐼𝜆,𝑒+𝑟 (𝜆, 𝜃, 𝜙) by the expression:

2𝜋 2𝜋
𝐽𝜆 (𝜆) = ∫0 ∫0 𝐼𝜆,𝑒+𝑟 (𝜆, 𝜃, 𝜙)𝑐𝑜𝑠𝜃𝑠𝑖𝑛𝜃𝑑𝜃𝑑𝜙 (2.9)

Figure 2.5. Graphical representation of surface radiosity [12].

In general, all incoming radiation must be reflected, absorbed or transmitted according to:

𝜌𝜆 + 𝛼𝜆 + 𝜏𝜆 = 1 (2.10)

where for heat transfer applications, the dependence on incoming direction for absorptivity (as
well as reflectivity and transmissivity) and outgoing direction for emissivity is generally weak and
is commonly neglected, where it is assumed that the surface absorbs and emits diffusely.
Therefore, for an opaque surface for any given wavelength the following expression can be
applied:
𝜀𝜆 = 𝛼𝜆 = 1 − 𝜌𝜆 (2.11)

where the following definitions can be defined for the four fundamental radiative properties:

16
 𝑟𝑒𝑓𝑙𝑒𝑐𝑡𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛
Reflectivity: 𝜌= (2.12)
𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

𝑎𝑏𝑠𝑜𝑟𝑏𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

Absorptivity: 𝛼= (2.13)
𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

𝑒𝑚𝑖𝑡𝑡𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

Emissivity: 𝜀= (2.14)
𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

𝑡𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

Transmissivity: 𝜏= (2.15)
𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛

When assessing real surfaces, blackbody radiation serves as a standard against which radiative
properties of actual surfaces can be compared. Blackbody surfaces are perfect absorbers and
emitters where α = ε = 1, and where its emitted radiation is taken with respect to wavelength,
independent of direction. Cavities with small apertures can be closely approximated as a
blackbody, especially if its inner surface is at a uniform temperature where entering radiation is
likely to experience many reflections before it can reemerge back to the ambient.
For surfaces with a prescribed temperature T, its emissivity is defined as the ratio of the
radiation emitted by the surface to the radiation emitted by a blackbody at the same temperature.
With respect to wavelength and direction, this parameter may be defined in terms of radiation
intensity by:
𝐼 (𝜆,𝜃,𝜙,𝑇)
𝜀𝜆,𝜃 (𝜆, 𝜃, 𝜙, 𝑇) = 𝜆,𝑒 (𝜆,𝑇) (2.16)
𝐼𝜆,𝑏

For inclined surfaces, Fig. 2.6, the angle of incidence for surfaces sloped due north or due south
can be determined with respect to the angle of inclination β, which has the same angular
relationship to direct beam radiation as a horizontal surface does at an artificial latitude of 𝜙 − 𝛽,
where 𝜙 is the respective latitude. Therefore, the angle of incidence can be determined for the
northern and southern hemispheres respectively by:

𝑐𝑜𝑠𝜃 = 𝑐𝑜𝑠(𝜙 − 𝛽)𝑐𝑜𝑠𝛿𝑐𝑜𝑠𝜔 + 𝑠𝑖𝑛(𝜙 − 𝛽)𝑠𝑖𝑛𝛿 (2.17)

𝑐𝑜𝑠𝜃 = 𝑐𝑜𝑠(𝜙 + 𝛽)𝑐𝑜𝑠𝛿𝑐𝑜𝑠𝜔 + 𝑠𝑖𝑛(𝜙 + 𝛽)𝑠𝑖𝑛𝛿 (2.18)

A more detailed description of these derived equations can be found by [14]. As prescribed by
Duffie and Beckman [14] it is possible to increase the radiation incident on an absorber by use of
planar reflectors, where the amount of radiation received from the reflector is a function of the
view factor. A view factor is traditionally defined as the fraction Frc of the radiation leaving a
reflective surface Ar that is intercepted by a collector surface Ac. A reciprocity expression expressed
by Eqn. 2.19 is also useful in determining one view factor from another.

𝐴𝑟 𝐹𝑟𝑐 = 𝐴𝑐 𝐹𝑐𝑟 (2.19)

17
Figure 2.6. Solar collector reference angles with respect to the equator with respect to a northern
hemisphere orientation [14].

As shown in Fig. 2.7, if one considers these two intersecting planes with an angle between the
planes ψ, then the summation of the view factors must equal 1, as prescribed by Eqn. 2.20.

𝐹𝑟𝑠 + 𝐹𝑟𝑐 + 𝐹𝑟𝑔 = 1 (2.20)

where s and g represents the projected sky and ground respectively, and where n and p represent
the respective lengths of the collector and reflector.

Figure 2.7. Cross-section of a collector and reflector [14].

Finally, in general for most engineering heat transfer applications Eqn. 2.21 is employed to
characterize the radiative heat transfer interaction between an object and its environment with
respect to wavelength.
𝑄̇𝑟𝑎𝑑 = 𝜎𝜀𝜆 𝐴(𝑇 4 − 𝑇∞4 ) (2.21)

where 𝜎 is the Stefan-Boltzmann constant, 5.67x10-8 W/m2-K4, A is the object surface area and 𝑇∞
is the ambient temperature.

18
2.2 Conduction Heat Transfer

Thermal conduction heat transfer through solids or stationary fluids occurs by a combination of
two physical mechanisms: lattice vibration and particle collision. Heat is conducted through
stationery fluids primarily by molecular collisions while in solids atoms are bound to each other
by a series of molecular bonds. When there is a temperature difference in a solid, the hot side of
the solid experiences more vigorous atomic movements which transfer thermal energy through
vibrational motion. Eventually both hot and cold sides reach thermal equilibrium, where all atoms
are vibrating with the same energy. Dense solids, especially metals have free electrons and are not
bound to any particular atom where they can freely move. The electrons within the hot side of a
solid will tend to move faster than those on the cooler side, and as the electrons undergo a series
of collisions, faster electrons give off their energy to slower ones. Eventually, after many
progressive collisions, equilibrium is reached where the majority of all electrons are moving at the
same average velocity. Conduction through electron collision is more effective than by lattice
vibration, which justifies why metals are generally better thermal conductors than ceramic
materials, which do not have many free electrons [15]. In fluids the conduction mechanism is
similar to metals where collisions between freely moving molecules facilitate heat transfer. The
effectiveness by which heat is transferred through a material is measured by its respective thermal
conductivity, k. A good conductor, such as copper, has a high thermal conductivity, whereas a
poor conductor (an insulator) has a low thermal conductivity. The rate of heat transfer by
conduction is given by Fourier’s Law:
𝑑𝑇
𝑄̇𝑐𝑜𝑛𝑑 = 𝑘𝐴 𝑑𝑥 (2.22)

where A is the cross-sectional area through which the heat is conducting, ∆T is the temperature
difference between two surfaces separated by a distance dx. In heat transfer, a positive q indicates
that heat is flowing into the body, and a negative q represents heat leaving the body.
For many CSP applications, heat transfer can be time dependent in nature for which
unsteady or transient initial and boundary conditions apply. If the surface temperature of a system
is altered, the temperature at each point in the system will also begin to change until a steady-state
temperature distribution is reached. In certain situations, temperature gradients within a solid
object can be approximately neglected where a lumped capacitance method may be applied to
determine the variation of temperature with respect to time. For simple situations for which sudden
changes in an environment facilitates transient conduction through an object initially at a uniform
initial temperature Ti to a lower surrounding fluid environment at 𝑇∞ the temperature of the solid
will decrease for time t > 0 until it reaches the ambient temperature. This reduction is due to
convection at the solid-liquid interface, where the lumped capacitance method can be utilized as
long as the solid is spatially uniform at any instant during the transient process, but only for which
the approximation that temperature gradients though the solid are negligible. To determine the
time required for a solid to reach a steady state temperature T, Eqn. 2.23 can be used in this context
of lumped capacitance.
𝑇−𝑇∞ ℎ𝐴
= 𝑒𝑥𝑝 (− 𝜌𝑉𝐶𝑠 𝑡) (2.23)
𝑇𝑖 −𝑇∞ 𝑝

where the thermal time constant τ can be considered according to:

19
 𝜌𝑉𝐶𝑝
𝜏= = 𝑅𝐶 (2.24)
ℎ𝐴𝑠

such that R is the resistance to convection and C is the lumped thermal capacitance of the solid,
which can slow the response of the solid to its thermal environment as these values increase,
analogous to an RC electrical circuit where a capacitive discharge through a resistor can facilitate
a voltage decay. To determine if the lumped capacitance approximation is valid, the Biot number,
Bi can be computed according to the form:

ℎ𝐿𝑐 𝑅𝑐𝑜𝑛𝑑 𝐿𝑐 /𝑘𝐴 𝐶𝑜𝑛𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑇ℎ𝑒𝑟𝑚𝑎𝑙 𝑅𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑇𝑠,1 −𝑇𝑠,2

𝐵𝑖 = = = = = (2.25)
𝑘 𝑅𝑐𝑜𝑛𝑣 1/ℎ𝐴 𝐶𝑜𝑛𝑣𝑒𝑐𝑡𝑖𝑜𝑛 𝑇ℎ𝑒𝑟𝑚𝑎𝑙 𝑅𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑇𝑠,2 −𝑇∞

If the Biot number is less than 0.1 for a solid object, then the entire material will be nearly the
same temperature with the dominant temperature difference will be at the surface where it can be
considered to be thermally thin. The Biot number must generally be less than 0.1 to use the lumped
capacitance approximation. This mathematical solution to the lumped system approximation
provides the basis for Newton's law of cooling [12]. Finally, with the definition of the Bi number
provided by Eqn. 2.25, rearrangement of Eqn. 2.23 provides Eqn. 2.26.
𝑇−𝑇∞
= 𝑒𝑥𝑝(−𝐵𝑖 ∙ 𝐹𝑜) (2.26)
𝑇𝑖 −𝑇∞

Here the parameter, Fo is dimensionless time which characterizes transient conduction problems
that also has the form:
𝑘
( )∙𝑡
𝜌𝐶𝑝
𝐹𝑜 = 2 (2.27)
𝐿𝑐

where LC is the characteristic length which is defined according to its geometry, Table 2.1.

Table 2.1. Characteristic length based on geometry.

Rectangular Cylindrical Spherical

𝑉𝑅𝑒𝑐 𝐿3 𝑉𝐶𝑦𝑙 𝜋𝑟 2 𝑙 𝑟 𝑉𝑆𝑝ℎ𝑒𝑟𝑒 4⁄3 𝜋𝑟 3 𝑟

𝐿𝐶 = = =𝐿 𝐿𝐶 = = = 𝐿𝐶 = = =
𝐴𝑅𝑒𝑐 𝐿2 𝐴𝐶𝑦𝑙 2𝜋𝑟𝑙 2 𝐴𝑆𝑝ℎ𝑒𝑟𝑒 4𝜋𝑟 2 3

Under conditions for which temperature gradients aren’t negligible, where 1D heat transfer is
facilitated, exact heat equation solutions may be used to compute the temperature dependence with
respect to time and space. These solutions exist for “finite solids” (plane walls, long cylinders with
small wall thickness, and spheres) and semi-infinite solids [12].
In general, for good conductors with metallic bonds, electrons from each atom are not
bound to any particular atom where they are very loosely held as free electrons, and can move
throughout the metal material. Free electrons carry heat across a material very quickly through
rapid collisions, where one electron gives some of its energy to another. In contrast, insulators do
not collide as rapidly with each other and if the material is solid, then its electrons are going to be
tightly held via covalent or ionic bonds. Gases and porous solids with large air-containing voids

20
can make the best insulators. Table 2.2 provides a list of common materials with their respective
thermal conductivities.

Table 2.2. Approximate thermal conductivity and heat capacities of common industrial materials
at 25 ᵒC and 1 atm [16].

When heat is added to a system, particles gain kinetic energy and since their average kinetic energy
increases, the temperature increases. Since these particles have increased their kinetic energy, they
move faster and further, where each particle takes up more room, so the material as a whole
expands in all directions while the mean free path can grow.
Finally, it is important to design with respect to material expansion coefficient differences
between adjacent materials, as a function temperature to reduce the onset of mechanical stress.
There are effectively three types of thermal expansion that can occur during operation: linear, area
and volume. If an arbitrary linear rod is heated up or cooled down, the rod will expand or contract
respectively, where thermal linear expansion is defined by Eqn. 2.28:

21
 ∆𝐿 = 𝛼𝐿0 Δ𝑇 (2.28)

such that ΔL is the change in length, α is the coefficient of linear expansion, L0 is the original
length of the rod, and ΔT is the temperature difference across the material. The change in length
of a rod is directly proportional to the temperature difference in the material and the coefficient of
thermal linear expansion. For a heated surface and volume, the coefficient αL0 from Eqn. 2.28 is
converted to 2αA0 and 𝛽𝑉0 .

2.3 Convection Heat Transfer Systems

Convection heat transfer occurs between a fluid in motion and a bounded surface or secondary
fluid, when the two are at differing temperatures. It may be classified according to the nature of
the fluid flow where forced convection occurs due to external means such as a pump, a fan or
mechanical energy source that facilitates motion as opposed to natural convection (or free) which
is facilitated by buoyancy forces that arise from density differences. With natural convection fluid
movement is created by the warm fluid itself. The density of the fluid decreases as it is heated;
thus, hot fluids are lighter than those that are cooler. Here warm fluid surrounding a hot object then
rises and is replaced by the cooler fluid, resulting in circulation. Although systems can be analyzed
from a standpoint of natural or forced convection, often mixed convection conditions arise where
velocities associated with a flow are small and/or buoyancy forces are large, and a secondary flow
that is comparable to the imposed forced flow could be induced. The buoyancy-induced flow
would be normal to the forced flow and could have a significant effect on convection heat transfer.
Additionally, the impacts of mixed convection could also be represented by an external fan which
could force fluid upward which would assist buoyancy, whereby a downward forced convective
flow would hinder it. For these examples convective heat transfer is described by energy transfer
occurring within the fluid due to combined effects of fluid conduction and bulk fluid motion where
energy that is transferred is wither sensible or internal fluid energy. However, during phase change
processes between a liquid to a vapor, latent heat exchange energy is added or removed based on
the respective process of boiling or condensation. For any type of convection, the magnitude of
heat exchange can be measured by the heat exchange coefficient h, prescribed by the rate Eqn.
2.29:
𝑄̇𝐶𝑜𝑛𝑣 = ℎ𝐴𝑆 (𝑇𝑆 − 𝑇∞ ) (2.29)

where 𝑄̇𝐶𝑜𝑛𝑣 is the convective power, proportional to the difference between the heated surface TS
and the ambient temperature 𝑇∞ . According to Incropera and Dewitt [12] convection heat transfer
coefficients can be considered according to Table 2.3, where phase change processes can be many
orders of magnitude higher than any conduction process. This dramatic heat transfer disparity has
been exploited in CSP with heat pipes where various Dish Stirling investigations [17] have found
marked performance improvements over sensible systems with heat pipes, which operate based on
both boiling and condensation processes.

22
 Table 2.3. General convection heat transfer coefficient values [12].

Generally, for single-phase heat transfer, the Nusselt number correlation is represented by an
empirical expression of the form:
𝜇 0.14
𝑁𝑢 = 𝐶 ∙ 𝑅𝑒 𝑝 ∙ 𝑃𝑟 𝑛 (𝜇 ) (2.30)
𝑠

where C, p and n in are constants independent of the fluid used [18]. The last term in this expression
accounts for the variable viscosity effect. This parameter is equivalent to the heat transfer
coefficient by the expression:
ℎ∙𝐿
𝑁𝑢 = 𝑘 𝑐 (2.31)
𝑓

Where h is the convective heat transfer coefficient, Lc is the characteristic length and kf is the
thermal conductivity of the surrounding fluid. Modified Wilson plot technique is primarily used
to determine the value of multiplier C and exponent of Reynolds number p in the Nusselt number
correlation [19]. The magnitude of n is taken to be 1/3 in agreement with [20] and [21].
Many CSP applications involve internal flow applications with convection transport
through varying cross-sectional geometries. To a first approximation a hydraulic diameter, used to
determine Nu can be approximated as the characteristic length by:
4𝐴𝐶
𝐷ℎ = 𝐿𝐶 = (2.32)
𝑃

where AC and P are the flow cross-sectional area and the wetted perimeter respectively. For
turbulent flow conditions, Re ≥ 2300, the Dittus-Boelter Eqn. 2.33 can be used for circular tubes,
as well as the Petukhov model, Eqn. 2.33, used for flows that are characterized by large property
variations.
⁄
𝑁𝑢𝐷ℎ = 0.023𝑅𝑒𝐷4ℎ5 𝑃𝑟 𝑛 0.6 ≤ 𝑃𝑟 ≤ 160
𝑅𝑒 ≥ 10,000
𝑛 = 0.4 𝑓𝑜𝑟 𝑇𝑠 > 𝑇𝑚 𝑓𝑜𝑟 { 𝐷ℎ (2.33)
𝐿
𝑛 = 0.3 𝑓𝑜𝑟 𝑇𝑠 < 𝑇𝑚 ≥ 10
𝐷ℎ

(𝑓⁄8)(𝑅𝑒𝐷 −1000)𝑃𝑟 0.5 < 𝑃𝑟 < 2,000

𝑁𝑢𝐷ℎ = 1.07+12.7(𝑓⁄8ℎ)1⁄2 (𝑃𝑟2⁄3 −1) 𝑓𝑜𝑟 { (2.34)
104 < 𝑅𝑒𝐷ℎ < 5 × 106

For smaller Re numbers the Gnielinski modified correlation has been suggested [22].

23
 (𝑓⁄8)(𝑅𝑒𝐷 −1000)𝑃𝑟 0.5 < 𝑃𝑟 < 2,000
𝑁𝑢𝐷ℎ = 1.07+12.7(𝑓⁄8ℎ)1⁄2 (𝑃𝑟2⁄3 −1) 𝑓𝑜𝑟 { (2.35)
3,000 < 𝑅𝑒𝐷ℎ < 106

For liquid metals with constant Ts the following correlation can also be used for Pe numbers greater
than 100.
𝑁𝑢𝐷ℎ = 5.0 + 0.025𝑃𝑒𝐷0.8 ℎ
(2.36)

For internal laminar flow through a duct with a uniform heat flux, the use of Eqn. 2.36 for various
cross-sectional areas can be used to determine 𝑁𝑢𝐷ℎ . Conversely, external flow over solid surface
boundary layers develop freely, without constraints imposed by adjacent surfaces.

Table 2.4. Laminar flow Nu numbers and friction factors for fully developed flow in tubes of
various cross-sectional areas [12].

Laminar boundary layers develop at the leading edge (x = 0) and transition to turbulence at a
downstream location, xc when a critical Reynolds number, Rex,c is achieved [12]. Eqns. 2.39 and
2.40 provide respective laminar and turbulent correlations for flow over flat surfaces [12] where
the ratio between the hydrodynamic and thermal boundary layers can be provided by:

𝛿
= 𝑃𝑟 1⁄3 (2.37)
𝛿𝑡

5.0 5𝑥
where 𝛿= = 𝑅𝑒 (2.38)
√𝑢∞ ⁄𝑣𝑥 𝑥

⁄
Laminar: 𝑁𝑢𝐿𝐶 = 0.664𝑅𝑒𝐿1𝐶 2 𝑃𝑟 1⁄3 𝑓𝑜𝑟 {𝑃𝑟 ≥ 0.6
̅̅̅̅ (2.39)

24
 ⁄
𝑁𝑢𝐿𝐶 = (0.037𝑅𝑒𝐿4𝐶 5 − 𝜍)𝑃𝑟 1⁄3
̅̅̅̅ 0.6 ≤ 𝑃𝑟 ≤ 60
Turbulent: 4 5 ⁄ 1 2 ⁄
𝑓𝑜𝑟 { 𝑅𝑒𝑥,𝑐 ≤ 𝑅𝑒𝐿𝐶 ≤ 108 (2.40)
𝜍 = 0.037𝑅𝑒𝑥,𝑐 − 0.664𝑅𝑒𝑥,𝑐

For free falling liquid drops Gourdon et al. [23] also provided a correlation for the average Nusselt
number, Eqn. 2.41 where all fluid properties are evaluated at 𝑇∞ :

⁄
𝑁𝑢𝐿𝐶 = 2 + 0.6𝑅𝑒𝐿1𝐶 2 𝑃𝑟 1⁄3 (2.41)

Heat transfer to or from fins in cross-flows are relevant to numerous thermal dissipation, boiling
and heat exchanger applications in CSP. During heat exchange operation, a fluid can pass over a
single tube where heat transfer can vary according to the cross-section as prescribed by Table 2.5.

Table 2.5. Average Nusselt number correlation equations for flow over varying single tube
geometries [360].

For a bank of tubes, the geometries can either be solid or have a secondary fluid passing through
them. The tubes can also either be aligned or staggered to the flow velocity, as shown in Fig. 2.8.

a. b.
Figure 2.8. Tube bank arrangements for a. aligned and b. staggered configurations.

25
For the staggered configuration the tube diameter D, traverse pitch ST, and longitudinal pitch SL
characterizes the tubes, measured between their respective centers. Values for these and other tube
bank configuration performance parameters and theory can be found in [12], where correction
coefficients C and m are provided and used to determine Nu numbers under specific flow
conditions. For an airflow across a bank of tubes, with more than NL rows of tubes, the average
heat transfer coefficient can be determined from Eqn. 2.42 where all properties are evaluated at
the film temperature [12].
𝑁𝐿 ≥ 10
̅̅̅̅ 𝑚
𝑁𝑢𝐷 = 1.13𝐶1 𝑅𝑒𝐷,𝑚𝑎𝑥 𝑃𝑟 1⁄3
𝑓𝑜𝑟 {2,000 < 𝑅𝑒𝐷,𝑚𝑎𝑥 < 40,000 (2.42)
𝑃𝑟 ≥ 0.7

Eqn. 2.43 by Zhukauskas [24] found improved accuracy for tube rows greater than 20, where all
fluid properties with the exception of Prs, were evaluated at the arithmetic mean of the fluid inlet
and outlet temperatures.
1⁄4
𝑁𝐿 ≥ 20
̅̅̅̅ 𝑚 0.36 𝑃𝑟
𝑁𝑢𝜙 = 𝐶𝑅𝑒𝐷,𝑚𝑎𝑥 𝑃𝑟 (𝑃𝑟 ) 𝑓𝑜𝑟 {1,000 < 𝑅𝑒𝐷,𝑚𝑎𝑥 < 2 × 106 (2.43)
𝑠
0.7 < 𝑃𝑟 < 500

Finally as the fluid that passes through the tube bank it can experience a large temperature change,
the heat transfer rate could be significantly overpredicted using the simplified expression from
Newton’s Law of cooling: ∆𝑇 = 𝑇𝑠 − 𝑇∞ . As its temperature approaches Ts and ΔT decreases, the
log-mean temperature difference, Eqn. 2.44 has been found to have enhanced results [12], where
the outlet temperature can be predicted from Eqn. 2.45.

(𝑇𝑠 −𝑇∞ )−(𝑇𝑠 −𝑇𝑂 )

∆𝑇𝑙𝑚 = 𝑇𝑠 −𝑇𝑖 (2.44)
𝑙𝑛( )
𝑇𝑠 −𝑇∞

𝑇𝑠 −𝑇𝑂 𝜋𝐷𝑁ℎ
= 𝑒𝑥𝑝 (− 𝜌𝑉𝑁 ) (2.45)
𝑇𝑠 −𝑇𝑖 𝑇 𝑆𝑇 𝑐𝑝

With the calculation of ∆𝑇𝑙𝑚 the heat transfer rate per unit length of the tubes within the bank can
be determined by Eqn. 2.46 where N is the total number of tubes and ℎ is the average heat transfer
coefficient.
𝑄̇ ′ = 𝑁(ℎ𝜋𝐷∆𝑇𝑙𝑚 ) (2.46)

2.4 Phase Change & Latent Heat Transfer

To provide electricity from CSP energy generation during off hours, latent HTFs and latent heat
thermal energy storage (LHTES), using phase change materials (PCMs), is receiving considerable
attention. Salts and other metallic-type materials are good candidates for utility as a PCM of high
temperature facilities, although charging and discharging power of LHTES systems can be limited
based on their respective thermal conductivities. Principally, latent heat is the energy absorbed or
released from a substance during phase change from a gas to a liquid or a solid, or vice versa. If a
substance is changing from a solid to a liquid, it will need to absorb energy from the surrounding
environment to spread the molecules into a larger volume. If the substance is changing however

26
from matter with lower density to a phase with higher density, the substance therefore will give
off energy as the molecules come closer together and lose energy from motion and vibration. As a
liquid undergoes boiling, energy is absorbed from the heated surface and goes into expanding the
liquid molecules into a vapor. Alternatively, when the liquid undergoes condensation, the liquid
gives off energy as it solidifies. Fig. 2.9 provides a simplified graphical example for how this
takes place, where a substance is in its solid phase at point 1 and is heated until its temperature
reaches its melting point at point 2. As it is heated further between points 2-3, the energy from the
heat source goes into breaking atomic bonds. At point 3 all of the solid phase has been transformed
into the liquid phase and further heat input goes into kinetic energy of the particles raising the
temperature until boiling, point 4. From points 4 to 5 thermal energy overcomes chemical bonds
to allow the particles that have enough kinetic energy to escape from the liquid where the substance
then enters into a gas phase. Beyond point 5, further heating under increased pressure can raise the
temperature further. With sensible heat however, when an object is heated or cooled, its
temperature rises or falls with the respective addition or removal of heat. However, the object does
not undergo phase change.

Figure 2.9. Phase change diagram based on temperature and input power.

During boiling, according to Carey [25] three mechanisms play important roles in boiling
processes: 1. surface tension effects, 2. surface wetting characteristics of the liquid, and 3.
metastable phase stability.

2.4.1 Surface Tension & Boiling

Surface tension σ corresponds to energy stored in the interface region per unit area [25]. The energy
excess in this region is due to the slightly larger separation of the liquid phase molecules adjacent
to the gas phase. The magnitude of the surface tension for a substance is directly linked to the
strength of intermolecular forces in the material. Nonpolar liquids typically have the lowest surface
tension. Water and other polar molecules have somewhat higher surface tension, and liquid metals,
which exhibit metallic bond attraction, have very high surface tension. The surface tension of water
at 20 °C is 0.0728 N/m, whereas liquid mercury has a surface tension of 0.484 N/m at the same
temperature. The surface tension for any pure liquid varies with temperature. It decreases almost
linearly with increasing temperature, vanishing altogether at the critical point where the distinction
between the phases disappears [25]. Therefore to ensure bubble interfacial stability, the pressure
inside a spherical bubble of radius r must exceed that in the surrounding liquid by2𝜎⁄𝑟:

27
 2𝜎
𝑃𝑣 − 𝑃𝑙 + (2.47)
𝑟

To ensure the bubble is in equilibrium with the surrounding fluid, the liquid must be superheated
above the respective saturation temperature for the ambient liquid pressure where the amount of
required superheating increases as the radius of curvature of the bubble interface decreases [25].

2.4.2 Wetting Characteristics

The wetting characteristics of a heat transfer fluid is generally quantified based on a contact angle
between the solid heated surface and the tangent to the interface, at the point where it contacts the
solid [25]. This angle is measured through the liquid phase (Fig. 2.10a) where it is generally
determined based on liquid advancing over a solid surface, which tends to be larger than that of a
liquid front receding over the surface, which is commonly termed as contact angle hysteresis Fig.
2.10b. The wetting behavior in combination with the surface tension effect, thus, determines the
level of superheat required for the bubble to be in equilibrium with the surrounding liquid. The
liquid must be heated above this superheat level for the bubble to grow. A steady boiling process
can be sustained only if the liquid is heated above this threshold superheat level [25].

a. b.
Figure 2.10. Contact angle θ a. comparison between hydrophobic and hydrophilic surfaces [27]
and b. hysteresis phenomena [28].

2.4.3 Phase Stability

It can be shown from basic thermodynamic analysis that a necessary and sufficient condition for
phase stability where:
𝜕𝑃
(𝜕𝑣 ) < 0 (2.48)
𝑟

Below the critical temperature, extrapolation of the isotherms for the liquid and vapor phases
consistent with an equation of state like the van de Waals equation results in an isotherm shape
similar to that shown in Fig 2.11.

28
 Figure 2.11. P-V diagram spinodal lines and metastable regions of boiling stability [25].

From Fig. 2.11 the locus of points where (𝜕𝑃⁄𝜕𝑣) 𝑇 = 0 are termed spinodal curves. Regions of
metastable vapor and liquid exist between the saturation curve and the spinodal curves. The effects
of surface tension discussed above require that fluid surrounding a vapor bubble be in the
metastable superheated liquid region. Predictions of statistical thermodynamics imply that as
(𝜕𝑃⁄𝜕𝑣) 𝑇 approaches zero, the level of fluctuations in a fluid system increases. This, in turn,
increases the probability that an embryonic new phase will form as a result of density fluctuations.
Initiation of a phase change in this manner is termed homogeneous nucleation. Generally, a pure
liquid must be heated to nearly 90% of its absolute critical temperature before homogeneous
nucleation of vapor bubbles occurs [25]. In most physical systems of engineering interest, the bulk
phase is in contact with solid walls of the containing structures, or solid particulate contaminants.
These solid phases may provide nucleation sites where phase change may occur if the system state
is driven into the metastable range. Nucleation of vapor bubbles may preferentially occur at low
liquid superheat levels in crevices in the solid surface where gas is trapped [25].
Overall, phase transition from a liquid to a vapor state strongly depends on temperature
and pressure, however forces, as a result of fluid motion can also impact bubble nucleation and
stability. In general, although forced convection and pool boiling can prescribe liquid to vapor
phase transition and resulting the appearance of vapor bubbles on a hot surface, forced convection
however imparts a number of specific features to the conditions of bubble production and
breakaway into the bulk of the liquid. The two-phase structures of vapor-liquid mixtures resulting
from boiling, and mixing of liquid and vapor phases, also differ appreciably between pool and
forced convection boiling. However, many leading investigators have previously [Kandlikar
paper] approached characterization of forced convection boiling to include physical phenomena
of pool boiling. Therefore, this work provides an overview of both boiling concepts, which are
central elements to latent phase-change heat transfer in CSP systems.

29
2.4.4 Pool & Convective Boiling Heat Transfer

2.4.4.1 Pool Boiling Heat Transfer

For CSP systems, much sensible heating and storage takes place for general HTF operation of oil-
based systems below 400 ᵒC and 550 ᵒC for molten salt systems [28], where Fig. 2.12 presents the
relative range for these and other CSP HTFs.

Figure 2.12. CSP heat transfer fluid relative operating range [139].

However, beyond these temperatures, phase change begins to occur according to the generalized
boiling curve, shown in Fig. 2.1.3. This curve primarily represents the points beyond point 3 in
Fig. 2.4.1, where initially natural convection heat transfer takes place between the heated surface
and the fluid up to point A in Fig. 2.42. In this region the superheat ΔT is very low where there is
insufficient vapor in contact with the liquid phase to facilitate boiling at saturation temperature.
However, as superheat is increased, bubble inception will eventually begin at the ONB (onset of
nucleate boiling) condition at point A. For the nucleate boiling region between ΔTA≤ ΔT ≤ ΔTC two
different flow regimes take place where first isolated bubbles form at nucleation sites, and as the
surface if further heated these bubbles separate from the heated surface at an increased rate, which
can cause considerable mixing and can enhanced the heat transfer coefficient h. As the superheat
is further increased beyond point B, more nucleation sites become active with increased bubble
formation causes bubble interference and coalescence. The heat flux in this region can be described
by correlations of nucleate pool boiling heat transfer where data have typically been used as tools
to predict nucleate boiling heat transfer in engineering systems and heat exchangers. A common
correlation used by many investigators in various applications is Eqn. 2.49, developed by
Rohsenow [29],
" 1⁄2 1⁄𝑟 1⁄𝑟
𝑄̇𝑁𝐵 𝜎 1 ⁄ 𝐶𝑝𝑙 [𝑇𝑤 −𝑇𝑠𝑎𝑡 (𝑃𝑙 )]
[𝑔(𝜌 −𝜌 )] = (𝐶 ) 𝑃𝑟𝑙−𝑠 𝑟 [ ] (2.49)
𝜇𝑙 ℎ𝑓𝑔 𝑙 𝑣 𝑠𝑓 ℎ𝑓𝑔

where the values of r = 0.33 and s = 1.7 are recommended for most fluids, though for water a value
of s = 1.0 is suggested. The values of Csf in this correlation vary with the type of solid surface and
the type of fluid in the system. This empirically accounts for material property and/or wetting angle
effects. Recommended values of Csf for specific liquid–solid combinations are given by Rohsenow
[29], but whenever possible, an experiment should be conducted to determine the appropriate value

30
of Csf for the particular solid–liquid combination of interest. If this is not possible, a value of Csf =
0.013 is recommended as a first approximation. Another common correlation for nucleate boiling
is Eqn. 2.50 by Rohsenow [29] where the impacts of surface-fluid interactions are considered, such
that surface-fluid and n coefficients provided for some materials are shown in Table 2.6.

1⁄2 3
𝑔(𝜌 −𝜌 ) 𝑐𝑝,𝑙 Δ𝑇𝑒
𝑄̇𝑁𝐵
"
= 𝜇𝑙 ℎ𝑓𝑔 [ 𝑙𝜎 𝑣 ] (𝐶 𝑛 ) (2.50)
𝑠,𝑓 ℎ𝑓,𝑔 𝑃𝑟𝑙

Table 2.6. Nucleate boiling correlation surface-fluid configuration coefficients [29].

This condition allows vapor to escape as jets or columns, which then form into larger slugs that
can blanket the heated surface. As this interfacial vapor blanketing increases, interference between
densely packed bubbles inhibits further liquid motion near the surface, which leads to an inflection
prescribing a heat transfer coefficient maximum, after which h begins to decrease with increasing
ΔT to the critical heat flux (CHF) condition where a further increase in superheat makes it difficult
for liquid to continuously wet the heated surface facilitating a dryout condition. For water/steam
systems, the CHF condition slightly exceeds 1 MW/m2.

Figure 2.13. Generalized pool boiling curve for HTFs [30]

31
Since high heat transfer rates and convection coefficients are associated with small values of excess
temperature, it is desirable to operate many types of CSP heat transfer equipment within the
nucleate boiling region [12]. For water/steam systems, dividing q” by ΔT demonstrates that
convection coefficients are higher (>104 W/m2-K) than other boiling regions. These high heat
transfer coefficients, especially for liquid metals (Add Table of Superheat, CHF and heat transfer
coefficients for water/steam, oils and metals here), is what makes latent energy systems for high
temperatures in excess of 550 ᵒC considerably attractive when compared to sensible systems [28].
To determine the CHF condition, Carey [25] prescribes Eqn. 2.51 where C is 0.149 for large
horizontal plates, and 0.131 for large horizontal cylinders, spheres and large finite heated surfaces.

𝜎𝑔(𝜌 −𝜌 )
𝑄̇𝐶𝐻𝐹 = 𝐶ℎ𝑙𝑣 𝜌𝑣 [ 𝜌𝑙2 𝑣 ] (2.51)
𝑣

After the CHF condition is reached, further superheat increase facilitates operation within the
transition boiling region. Here, unstable film or partial film boiling ensues, where bubble formation
occurs so rapidly that a vapor film blankets the surface to where at any point on the surface,
conditions may fluctuate between film and nucleate boiling. Since the thermal conductivity of the
vapor is less than that of the liquid, h and q” continue to decrease with increasing ΔT where
radiation heat transfer begins to become a more significant mode of heat transfer; however thermal-
mechanical stress and fatigue of the surface becomes more problematic. It is advised that operation
occur only up to the CHF condition to avoid system failures.
Boiling heat transfer, which can facilitate the highest heat transfer coefficients as illustrated by
Table 2.3, can either occur with homogeneous or heterogeneous nucleation, where homogeneous
occurs in the pure liquid state when a sufficient number of high-energy molecules come together
to form a void, into which a vapor immediately diffuses to for a nucleus [31]. Conversely,
heterogeneous nucleation occurs in the presence of two or three phases, which tends to occur more
frequently considering homogeneous nucleation requires a relatively large amount of energy
required to form the bubble surface and provide the work of expansion necessary to facilitate
bubble growth up to the critical bubble radius. For liquid metals, as with ordinary liquids, bubble
nucleation for incipient boiling occurs at gas-filled cavities in wetted solid-metal surfaces.
Additionally, for any surface to boil, the heated surface temperature must be higher than the
saturation temperature, where this difference is known as wall superheat. Whereas stable-nucleate
boiling superheats for liquid metals are typically less than those for ordinary fluids, incipient-
boiling wall superheats are often much greater. This indicates that for liquid metals wall
temperature may have a large overshoot at boiling inception with a significant amount of
hysteresis, where this overshoot is generally higher for pool boiling than forced convection boiling
[31]. Dweyer indicated [31] that there are four reasons why liquid metals generally provide much
higher incipient boiling superheats than with ordinary liquids:

1. Liquid metals usually wet solid metal surfaces very well, particularly in the range of
temperatures encountered in boiling.
2. Alkali metals are very reactive chemically and act to reduce surface oxides of many metals and
alloys.
3. Solubilities of inert gases in liquid metals increase with an increasing temperature.
4. Boiling pressures of liquid metals are generally low with respect to their critical pressures, which
means that boiling occurs in that portion of the vapor pressure curve where the slope, dp/dt is
low.

32
For pool boiling studies with alkali metals, Subbotin et al. [32] reported incipient-boiling
superheats of Sodium up to approximately 90 ᵒC, while Edwards and Hoffman [33] reported
incipient-boiling superheats as high as 260 ᵒC for potassium. For forced convection, Chen [34]
observed incipient-boiling wall superheats over 37 ᵒC with potassium, while Dweyer [32] found
superheats up to 126 ᵒC for sodium, with HTF flow faster than 0.3 m/s. Overall, the magnitude of
incipient-boiling superheats in metals has been found [31] to be impacted by the following
independent variables:

1. HTF operating pressure and partial pressure of inert gas in surface cavities.
2. Heat flux and heating method.
3. Flow rate or Reynolds number [35].
4. Pressure-temperature history of the system.
5. Liquid metal charging procedure, which can affect the amount of inert gas that can be trapped
within surface cavities.
6. Inert gas and O2 concentration in HTF.
7. Topography or microstructure and surface roughness.
8. Extent of natural convection (pool boiling), impacted by the system geometry.
9. Rate of temperature rise up to boiling temperature.
10. Operational time and reliability of the system.

2.4.4.2 Convective Boiling Heat Transfer

In contrast to pool boiling which is primarily facilitated by buoyancy-driven flow, forced

convection boiling is due to forced bulk motion of the fluid, which can also include motion due to
buoyancy effects. Convective boiling may also occur based on external flow over heated
geometries [12], where vapor generation is not extensive and the liquid is subcooled, or with
internal duct flow, commonly termed as two-phase flow. This type of flow is characterized by
rapid evaporation of liquid to vapor in the flow direction.

Figure 2.14. Convective flow boiling along a horizontal channel with qualitative variation of the
heat transfer coefficient with respect to quality, under moderate wall superheat conditions [25].

33
Fig. 2.14 illustrates the vaporization process in a horizontal round tube where subcooled liquid
enters prior to transitioning into two-phase flow. The regimes encountered depend on the entrance
conditions and the thermal boundary conditions at the tube wall. At low quality the vaporization
process is dominated by nucleate boiling, with convective effects being relatively weak. As the
quality increases, the flow quickly enters the annular film flow regime in which convective
evaporation of the annular liquid film is the dominant heat transfer mechanism.
At high wall superheat levels, transition boiling or film boiling can also occur. The
transition from nucleate boiling to one of these regimes is termed a departure from nucleate boiling
(DNB) or the CHF condition. However, the heat transfer performance of an evaporator under
transition or film boiling conditions is so poor that equipment is not usually designed to operate
under such conditions.
For convective flow boiling along a vertical channel, the impact of buoyancy can become
significant for bubble departure from the heated surface which has an impact on regime transition
with respect to the fluid. Fig. 2.15 presents convective boiling along a vertical channel where heat
transfer to the entering subcooled liquid can be predicted by convection correlations presented in
section 2.3 up to point B.

Figure 2.15. Flow convection heat transfer regions in a vertical tube [36].

In forced convection boiling, transitions into separate regimes occur as nucleate, bubbly, slug,
annular and dryout which correspond to flow of increasing quality. Each of these regimes have

34
separate experimental correlations associated to them where their transition from one to another is
typically not well defined. Bubble growth/separation and boiling regime transitions are strongly
affected by flow velocity and other hydrodynamic effects [12]. When boiling is initiated, bubbles
begin to nucleate on the heated surface and depart when a critical radius is achieved [25]. For this
regime, the wall heat flux is predicted by Carey [25] for a constant heat flux, Eqn. 2.52, and
isothermal boundary condition, Eqn. 2.53, where G is the Gibbs function: G = H-TS and x is a
prescribed position along the channel.

𝐷
Const. Heat Flux B.C. 𝑄̇ ∥ = ( 4𝑧ℎ ) 𝐺𝑐𝑝,𝑙 [𝑇𝑙 (𝑧) − 𝑇𝑙,𝑖𝑛 ] (2.52)

4ℎ 𝑧
Isothermal B.C. 𝑄̇ ∥ (𝑧) = ℎ𝑙 [(𝑇𝑤 − 𝑇𝑠𝑎𝑡 ) − (𝑇𝑠𝑎𝑡 − 𝑇𝑙,𝑖𝑛 )] exp (− 𝐺𝑐 𝑙𝐷 ) (2.53)
𝑝,𝑙 ℎ

As the volume fraction of vapor increases, nucleate boiling is observed at low vapor qualities
(x<0.1) which increases as individual bubbles coalesce to form slugs, which then proceed into an
annular flow with increasing heat input. As this occurs saturated nucleate boiling becomes
dominant where various investigators have proposed correlations to describe these phenomena
[37-39]. Fig. 2.16 presents the boiling curve for flow convection where the heat flux is plotted
verses the wall superheat, where the transition from forced convection to nucleate boiling is always
at some wall temperature greater than the saturation temperature. The exact point where the boiling
curve transitions from single phase to two-phase is dependent upon the amount of subcooling,
mass flow rate, number and size of nucleation sites, and the fluid parameters. Most expressions for
the prediction of ONB (onset of nucleate boiling) are derived by first determining the intersection
of the forced convection (line 1-2) and fully developed boiling portions (line 3a-3b) of Fig. 2.16.

Figure 2.16. Convective Boiling curve at onset of nucleate boiling [25].

Carey [25] suggests for various fluids high accuracy can be attained with a heat transfer correlation
by Kandlikar [40], based on either dominant contributions from nucleate boiling or convective
boiling:

35
 ℎ
ℎ𝑠𝑎𝑡 = 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑜𝑓 { 𝑁𝐵𝐷 (2.54)
ℎ𝐶𝐵𝐷

where for the nucleate boiling dominant regime:

𝜌 0.1
ℎ𝑁𝐵𝐷 = 0.6683 (𝜌 𝑙 ) 𝑥 0.16 (1 − 𝑥)0.64 𝑓2 (𝐹𝑟𝑙 )ℎ𝑙 + 1058.0𝐵𝑜0.7 𝐹𝐾 (1 − 𝑥)0.8 ℎ𝑙 (2.55)
𝑣

and for the convective dominant regime:

𝜌 0.45
ℎ𝐶𝐵𝐷 = 1.1360 (𝜌 𝑙 ) 𝑥 0.72 (1 − 𝑥)0.08 𝑓2 (𝐹𝑟𝑙 )ℎ𝑙 + 667.2𝐵𝑜0.7 𝐹𝐾 (1 − 𝑥)0.8 ℎ𝑙 (2.56)
𝑣

where FK is a fluid constant provided by Kandlikar [41], which is 1 for water and 4.7 for nitrogen.
In Eqn. 2.56 ℎ𝑙 is the single-phase heat transfer coefficient for the liquid phase only, and is
computed according to correlations by Gnielinski [22]:

𝑘 (𝑅𝑒𝑙 −1000)𝑃𝑟𝑙 (𝑓⁄2)

ℎ𝑙 = ( 𝐷𝑙) 2⁄3 𝑓𝑜𝑟 0.5 ≤ 𝑃𝑟𝑙 ≤ 200, 2300 ≤ 𝑅𝑒𝑙 < 104
1+12.7(𝑃𝑟𝑙 −1)(𝑓⁄2)0.5
𝑘 𝑅𝑒𝑙 𝑃𝑟𝑙 (𝑓⁄2)
(2.57)
ℎ𝑙 = ( 𝐷𝑙) 2⁄3 𝑓𝑜𝑟 0.5 ≤ 𝑃𝑟𝑙 ≤ 200, 104 ≤ 𝑅𝑒𝑙 < 5 × 106
1.07+12.7(𝑃𝑟𝑙 −1)(𝑓⁄2)0.5

From the aforementioned relations, the friction factor f and function f2 are computed according to:

𝑓 = [1.58𝑙𝑛(𝑅𝑒𝑙 ) − 3.28]−2 (2.58)

𝐺𝐷
𝑅𝑒𝑙 = (2.59)
𝜇𝑙

(25𝐹𝑟𝑙 )0.3 𝑓𝑜𝑟 𝐹𝑟𝑙 < 0.04 𝑓𝑜𝑟 ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑡𝑢𝑏𝑒𝑠

𝑓2 (𝐹𝑟𝑙 ) = { (2.60)
1 𝑓𝑜𝑟 𝐹𝑟𝑙 > 0.04 𝑓𝑜𝑟 ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑎𝑛𝑑 𝑣𝑒𝑟𝑡𝑖𝑐𝑎𝑙 𝑡𝑢𝑏𝑒𝑠

As with pool boiling, dry spots eventually appear on the inner surface, at which point the
convection coefficient begins to decrease at the CHF condition. The reader is referred to [42-44]
and CHF correlations by [25] for subcooled boiling and to [45-47] for CHF correlations associated
with saturated flow boiling.

2.4.5 CSP Multimode Heat Transfer

During operation, most CSP components within a system experience more than one form of heat
transfer where conduction, convection or radiation effects within an energy balance must be
considered respectively. As illustrated in Fig. 2.17 for a generalized CSP receiver, all three effects
can be evident within a prescribed control volume, where heat transfer analysis can be facilitated
by an arrangement of the continuity, momentum and energy balance constitutive equations. As
previously described, conduction assesses the diffusion of heat due to temperature gradients while
convection impacts are facilitated by fluid flow heat transfer. However, for systems at high

36
temperature, radiation effects, which is the transfer of energy by electromagnetic waves between
surfaces at differing temperatures, scale to the fourth power of temperature. This can have a
tremendous impact on CSP systems, especially heat losses [434]. Additionally, although the heat
transfer coefficient magnitude of these effects, as illustrated by Table 2.3, can vary widely
depending on control volume geometry and solid/fluid material properties, conjugate heat transfer
behavior for temperatures above 720 °C (DOE SunShot target temperature [8]) and pressures
above 1 MPa are still not well understood with regards to heat losses for many CSP technologies.

Figure 2.17. Multimode heat transfer of a generalized channeled receiver.

Much research is still ongoing to improve conjugate heat transfer performance and reliability of
CSP receivers and other high-temperature systems including thermal energy storage (TES) and
power block heat exchangers. The next subsequent chapters provide an overview of progress in
CSP systems from the standpoint of heat transfer enhancement.

37
3. THERMAL COLLECTORS
3.1 Overview

Solar thermal concentrating receivers are designed to be a high-absorptance/transmittance, low

reflectance, radiative/convective heat exchanger that emulates as closely as possible a radiative
black body. Many concentrator varieties are possible for increasing the flux of radiation for CSP
receivers. Duffie and Beckman [9] categorized these into two groups: reflectors and refractors,
whereby these can be designed cylindrical to focus on a “line” geometry or circular to focus onto
a “point.” Examples of six configurations are shown in Fig. 3.1 where receivers can also be
classified as concave, flat or convex. In general, concentrators with receivers much smaller than
the aperture are effective only on beam radiation. It is also evident that the angle of incidence of
the beam radiation on the concentrator is important and that sun tracking will be required for these
collectors [9].

a. b. c.

d. e. f.

Figure 3.1. Concentrating solar collector configurations for a. tubular absorbers with a diffuse
back reflector, b. tubular absorbers with specular cusp reflectors, c. plane receiver with plane
reflectors, d. parabolic concentrator, e. Fresnel reflector and f. array of heliostats with a central
receiver [9].

There are unique challenges associated with receivers for different CSP technologies that have
distinctive requirements to push the envelope of performance. These include the development of
geometric designs, high-temperature materials, HTFs as well as processes that maximize solar
irradiance and absorptance, minimize heat loss, and have high reliability at high temperatures over
thousands of thermal cycles [48]. Additionally, flat-plate collector systems are traditionally termed
“absorber plates”, while concentrating collectors are often referred to as “receivers.” For CSP
trough systems, the optical principle of a reflecting parabola is that all rays of light parallel to its
axis are reflected to a point or a linear-focus translation. A parabolic dish (paraboloid), on the other
hand, is formed by rotating the parabola about its axis; the focus remains a point and is often called
point-focus concentrators. If a receiver is mounted at the focus of a parabolic reflector, the
reflected light will be absorbed and converted into heat. These two principal functions, reflection
to a point or a line, and subsequent absorption by a receiver, constitute the basic functions of a
parabolic concentrating collector [48]. Depending on the optics design, certain receivers, such as

38
ominidirectional receivers, can achieve higher concentration ratios, and thus higher HTM
temperatures than that of closed cavity receivers, Fig. 3.2.

Figure 3.2. Variation of geometric concentration ratio with receiver rim angle [48].

As prescribed by Duffie and Beckman [9] the higher the temperature at which energy is to be
delivered, the higher must be the concentration ratio, and the more precise must be the optics of
both the concentrator and the orientation system. Fig. 3.3 shows practical ranges of concentration
ratios and types of optical systems needed to deliver energy at various temperatures where Aa and
Ar are the aperture and receiver areas respectively [9].

Figure 3.3. Concentration ratio and receiver temperature where the lower limit curve represents
concentration ratios where thermal losses equal absorbed energy, shaded range corresponds to
collection efficiencies of 40-60% of probable operation. Additionally, approximate ranges of three
classes of collectors are shown [49].

The heat-transfer media can be a liquid, such as water, synthetic oil, or molten salt; a gas/vapor,
such as air, helium or steam; a solid, such as sand or ceramic particles; or a supercritical phase,
such as supercritical carbon dioxide. Liquids and supercritical fluids are typically contained and
transported within a tubular receiver, whereas gases can be contained in tubes or volumetric

39
(honeycomb-like) receivers. Solid particles can be released into a cavity receiver for direct
irradiance or across irradiated tubes as a fluidized bed or moving packed bed. Current DOE
SunShot goals for a CSP receiver include exit temperatures ≥ 720 °C, thermal efficiency ≥ 90%,
lifetime ≥ 10,000 cycles and cost ≤ $150/kWth.
Receivers should be highly absorptive within solar wavelengths of approximately 0.4–2.5
μm. However, at high temperatures, receivers can emit a significant amount of thermal radiation,
thereby losing heat to the environment. At these high temperatures there is an overlap of the
spectral absorptance and emittance, making it difficult to develop a truly selective absorber.
Additionally, coatings and paints must endure high temperatures in an oxidizing environment for
many thermal cycles and under potentially large thermal gradients. Overall, the challenges [50]
that exist for improving the performance of high-temperature receivers are:

1. Development of optimized geometric designs.

2. High-temperature materials.
3. Heat-transfer fluids.
4. Maximize absorptance to minimize heat losses.
5. High reliability over thousands of thermal cycles.
6. Direct (no intermediate heat exchanger) vs. indirect (ability to store heat) heating.

3.2 Thermal Collectors Analysis

To perform an energy balance on a solar thermal collector, the incident surface is typically isolated
where the heated surface absorbs incoming radiation, and balances energy losses by:

𝑄̇𝑈𝑠𝑒𝑓𝑢𝑙 𝐿𝑜𝑠𝑠 = 𝐸̇𝑜𝑝𝑡,𝑖𝑛𝑐 − 𝐸̇𝑡ℎ 𝑙𝑜𝑠𝑠 (3.1)

where in general the losses can be evaluated based on an energy balance similar to the
representation provided in Fig. 3.4

Figure 3.4. Energy balance on a generalized solar absorber/collector [48].

40
For omnidirectional receivers, the glass envelope temperature Tg is assessed as a function of the
receiver tube temperature and the resultant rate of thermal energy exchange between the receiver
tube and glass envelop [48]. In his investigation, Treadwell [51] presented the following simplified
equation where the temperature of the glass envelope can be determined by evaluating the glass
heat loss rate, Eqn. 3.2 with the receiver tube heat-loss rate producing Eqn. 3.3.

𝑄̇𝑙𝑜𝑠𝑠,𝑔𝑙𝑎𝑠𝑠 = ℎ𝑔𝑙𝑎𝑠𝑠 𝐴𝑔𝑙𝑎𝑠𝑠 (𝑇𝑠𝑢𝑟𝑓

4 4
− 𝑇∞4 ) + 𝜎𝑆𝐵 𝜀𝑔𝑙𝑎𝑠𝑠 𝐹𝑟𝑎𝑑 𝐴𝑔𝑙𝑎𝑠𝑠 (𝑇𝑔𝑙𝑎𝑠𝑠 − 𝑇∞4 ) (3.2)

4 4
𝜎𝑆𝐵 𝐴𝑠𝑢𝑟𝑓 (𝑇𝑠𝑢𝑟𝑓 −𝑇𝑔𝑙𝑎𝑠𝑠 ) 2𝜋𝑘𝑒𝑓𝑓 𝑙𝑟𝑒𝑐 (𝑇𝑠𝑢𝑟𝑓 −𝑇𝑔𝑙𝑎𝑠𝑠 )
𝑄̇𝑙𝑜𝑠𝑠 = (1⁄𝜀 + (3.3)
𝑠𝑢𝑟𝑓 )+(𝐴𝑠𝑢𝑟𝑓 ⁄𝐴𝑔𝑙𝑎𝑠𝑠 )[(1⁄𝜀𝑔𝑙𝑎𝑠𝑠 )−1] 𝑙𝑛(𝑟2 ⁄𝑟1 )

where the subscripts “surf” and “rec” corresponds to the surface of the receiver tube and the
receiver respectively, while r1and r2 are the respective radial lengths in Fig. 3.5.

Figure 3.5. Omnidirectional metal tube receiver with outer glass tube diagram.

With Eqn. 3.3 a thorough derivation and values of ke can be found in reviews by [52-54], however
based on knowledge of kair Table 3.1 [48] provides values of ke/kair for various Ra numbers.
However, if an evacuated annulus receiver is employed then kair = 0.

Table 3.1. Ratio of effective conductance of annulus ke to air conductance with Rayleigh number
[48].

When performing an energy balance on the absorber plate of a solar collector, there are four
general important mechanisms that reduce the amount of solar energy that is incident on the

41
collector aperture: imperfect reflection, imperfect geometry, imperfect transmission and imperfect
absorption.
Based on a thermodynamic first law approach, Eqn. 3.4 provides a single expression for
the rate of useful energy produced from a solar collector, equivalent to the difference of the optical
(short wavelength λ) energy incident on the absorber surface and the respective heat losses by the
different modes of heat transfer [48].

𝑄̇𝑈𝑠𝑒𝑓𝑢𝑙 = 𝑚̇𝐶𝑝 (𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛 ) (3.4)

4
= Γ𝜌𝜏𝛼𝐼𝑖𝑛𝑐,𝑎 𝐴𝐴𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 − 𝐴𝑠 [ℎ𝑐𝑜𝑚𝑏 (𝑇𝑠 − 𝑇𝑎𝑚𝑏 ) + 𝜀𝜎(𝑇𝑠4 − 𝑇𝑠𝑘𝑦 )]

where hcomb is equal to the combined convection and conduction coefficient. Kalogirou [55]
derived an expression for the energy transfer to a fluid at a local fluid temperature Tf:

𝐴𝑎 𝑈𝐿
𝑞′𝑢𝑠𝑒𝑓𝑢𝑙 = 𝐹′ [𝑛0 𝐺𝑏 − (𝑇𝑓 − 𝑇𝑎 )] (3.5)
𝐿 𝐶

From Eqn. 3.5, the collector efficiency is defined as:

1⁄𝑈𝐿
𝐹′ = 1 𝐷0 𝐷 𝐷 (3.6)
+ +( 0 +𝑙𝑛 0 )
𝑈𝐿 ℎ𝑓𝑖 𝐷𝑖 2𝑘 𝐷𝑖

where D0 and Di are the riser tube outer and inner diameters, UL is the solar collector heat transfer
coefficient, L is the distance half distance between two consecutive riser pipes defined by [55]
where C is the concentration ratio, n0 is the collector optical efficiency and Gb is the direct beam
irradiation in units of (W/m2).
The rate of optical (short wavelength λ) radiation incident on an absorber/receiver will be
the solar irradiance absorption for that type of collector, also based on its direct and diffuse
components. Since the capture area of the collector may not be aimed directly at the sun, this
resource must be reduced to account for the angle of incidence [48]. The incident solar energy as
a function of the collector opening area (aperture) is therefore:

𝐸̇𝑖𝑛𝑐 = 𝐼𝑖𝑛𝑐,𝑎 𝐴𝐴𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 (3.7)

This solar resource encompasses a number of losses as it passes from the aperture of the collector
to the absorber. These processes depend on the type and design of the specific collector where
important optical loss mechanisms must also be included. The rate of optical (short wavelength λ)
energy reaching the absorber or receiver is the product of the incoming solar resource multiplied
by a number of factors, all less than 1.0:

𝐸𝑙𝑜𝑤𝜆 = Γ𝜌𝜏𝛼𝐼𝑖𝑛𝑐,𝑎 𝐴𝐴𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 (3.8)

where the capture fraction Γ is a parameter which represents the quality of the shape of the
reflecting surface and size of receiver, which is often described as the fraction of the reflected
energy not impinging on or entering the receiver or also termed “spillage.” If the receiver is
relatively small or the shape of the concentrator is poor, then this parameter will be considerably
less than 1.0. The reflectance, ρ is a property of any reflecting surface where well designed

42
concentrators can have capture fractions of 0.95 or higher [48]. The τ transmittance is the fraction
of solar radiation passing through all transparent cover material that sunlight passes through to the
absorber. The transmittance is the average overall value and represents the total reduction in
transmitted energy of the incident solar spectrum by all cover materials. Cover sheets of glass or
plastic are used on flat-plate collectors, above the absorber to reduce convective heat loss.
Absorptance 𝛼 and emittance 𝜀 are experimentally determined parameters, where radiation
properties of various surfaces are provided by Duffie and Beckman [9]:

Table 3.2. Radiation properties of various thermal collector materials [9].

High-temperature cavity revivers have been designed to employ quartz glass covers to reduce heat
loss, to separate the gas in the receiver from the outside air, or used for facilitating pressurization
of the gas within the cavity. The transmittance of the cover material also depends on the

43
wavelength of light passing through it where materials such as glass primarily transmits most
radiation in the visible spectrum, however does not transmit much in the infrared. Therefore, an
absorber covered with glass will receive most of the incoming, short wavelength radiation, while
not transmitting much of the long wavelength radiation loss coming from the absorber [48]. Plastic
however has high transmittance values at very long wavelengths, which tends to increase heat loss
during night time, lower-temperature ambient conditions [48].
Although a black surface can have an absorption coefficient value greater than 0.98 [48],
as the surface degrades this value can decrease. It is important to note that this property is for
radiation in the solar or “visible” spectrum where for most real surfaces, the absorption varies as a
function of the wavelength. This can be especially true for selective surfaces which are a class of
solar collector surfaces that have higher absorptance in the visible spectrum that reduces thermal
radiation losses [48]. As solar irradiance (short wavelength λ) reaches the surface of an absorber
it then raises the temperature above that of the ambient, which initiates heat loss from the absorber.
As the temperature difference between the collector surface and the ambient increases, this rate of
heat loss can impact the overall energy balance, Eqn. 3.9.

𝑄̇𝑙𝑜𝑠𝑠 = 𝑄̇𝑙𝑜𝑠𝑠,𝑐𝑜𝑛𝑣𝑒𝑐𝑡𝑖𝑜𝑛 + 𝑄̇𝑙𝑜𝑠𝑠,𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛 + 𝑄̇𝑙𝑜𝑠𝑠,𝑐𝑜𝑛𝑑𝑢𝑐𝑡𝑖𝑜𝑛 (3.9)

Adjusting the flow rate or having a large heated surface roughness can reduce the temperature of
the absorber, reducing heat loss. However, if enough heat is not removed from the absorber,
potentially due to interrupted flow or over insulated surface, the temperature of an absorber could
increase hazards to flow channel material stress and thermal fatigue [48].
With regard to convection, this loss from a solar collector receiver is proportional to the
surface area of the receiver and the temperature difference between the surface and ambient:

𝑄̇𝑙𝑜𝑠𝑠,𝑐𝑜𝑛𝑣𝑒𝑐𝑡𝑖𝑜𝑛 = ℎ̅𝑐 𝐴𝑐 (𝑇𝑐 − 𝑇𝑎𝑚𝑏 ) (3.10)

Although Eqn. 3.10 is a simplified representation, there are typically a number of convective and
radiative processes based on surface geometries and boundary layer development that influence
heat loss. The average temperature of a collector, Tc has uncertainty with its value as it’s not a
fixed or measureable quantity [48] since its temperature near the HTF inlet will be lower than near
the outlet, and both locations will be lower than the intermediate surfaces not in contact with the
HTF. Also, since convection is more of a surface phenomenon as it’s assessed with respect to
surface temperature, Tc may be driven by characteristics of a coating rather than the solid structural
material. Overall, since convective losses are a major loss term for solar collectors, CSP design
engineers tend to incorporate many features to collector designs to reduce its impact. Some of
these include:

1. Multiple transparent cover sheets for flat-plate collectors.

2. Glass tubes surrounding linear absorbers with a vacuum drawn in the intervening space.
3. Concentration of solar energy so that the absorber area is small relative to the capture area.
4. Absorbers within cavities incorporating glass windows.

Based on Eqn. 3.10 radiation losses increase as the ΔT above the ambient temperature grows. For
a black, vertical surface in still, ambient air at 25 ᵒC, Fig. 3.6 exhibits the impact of radiation losses
compared to convective losses, which is proportional to the surface emittance.

44
Figure 3.6. Convection versus Radiation Generalized Heat Losses for a Black, Vertical Surface
in Free Air at 25 ᵒC. [56].

Additionally, when determining an optimal size of a receiver many aforementioned factors of heat
and optical losses can impact the energy absorbed by a receiver. Fig 3.7 by Stine and Geyer [56]
illustrates that as the receiver size increases the amount of intercepted energy too increases as well
as heat losses. The sum of energy intercepted by the receiver and heat from the receiver will show
a maximum at an optimum receiver size [56].

Figure 3.7. Optimization of the collector receiver size [56].

Even with an optimized receiver, heat transfer can still be limited to other geometrical limitations
inherent to the type of system in operation. Fig. 3.8 provides daily energy performance data for
dish, trough and flat plate collectors, averaged over a year. It is shown that for the same operating
temperature of 288 ᵒC, receiver dish systems are capable of higher absorption than N/S or E/W
trough systems, for particular configurations, examined at Sandia National laboratories [48].

45
Figure 3.8. Annual performance summary of CSP collectors over a 1-year period at Sandia
National Laboratories [48].

For high-temperature systems, second law analysis can have an impact in assessing overall
performance based on entropy loss. An analysis by Kalogirou [55] based on a work by Bejan [57,
58] considered the maximization of power output based on the minimization of entropy, where the
exergy inflow from incident solar radiation falling on a collector is provided by:

𝑇
𝐸𝑥,𝑖𝑛 = 𝑄̇𝑁𝑒𝑡 (1 − 𝑇 0 ) (3.11)
𝑎𝑝𝑝

where 𝑄̇𝑁𝑒𝑡 is the net incident solar radiation on a collector, 𝑇𝑎𝑝𝑝 is the apperant sun temperature
as an exergy source, which according to Petela [59] is approximately equal to 3/4Ts, where Ts is
the apparent black body temperature of the sun at approximately 6000 K. Conversely the output
exergy from a collector is provided [55] as:

𝑇
𝐸𝑥,𝑜𝑢𝑡 = 𝑄̇ (1 − 𝑇0 ) (3.12)
𝑟

The difference between 𝐸𝑥,𝑖𝑛 and 𝐸𝑥,𝑜𝑢𝑡 represents destroyed exergy where the entropy generation
rate can be characterized by:
1
𝑆𝑔𝑒𝑛 = 𝑇 (𝐸𝑥,𝑖𝑛 − 𝐸𝑥,𝑜𝑢𝑡 ) (3.13)
0

𝑄̇0 𝑄̇ 𝑄̇
based on the expression: 𝑆𝑔𝑒𝑛 = + 𝑇 − 𝑇 𝑁𝑒𝑡 (3.14)
𝑇0 𝑟 𝑎𝑝𝑝

Eqn. 3.14 can also be re-written as 𝐸𝑥,𝑜𝑢𝑡 = 𝐸𝑥,𝑖𝑛 − 𝑇0 𝑆𝑔𝑒𝑛 where Fig. 3.9 illustrates the preceding
entropy equations where maximum performance can be achieved with a minimum value of 𝑆𝑔𝑒𝑛 .

46
 Figure 3.9. Exergy Analysis Flow Diagram [55]

Using Eqn. 3.14 in addition to a derived maximum dimensionless collector temperature 𝜃𝑚𝑎𝑥 [55]
Kalogirou determined an expression for the minimum dimensionless collector temperature, 𝜃𝑜𝑝𝑡
based on entropy minimization, Eqn. 3.15, which is derived based on Eqn. 3.16 [55]:

𝑞0∗ 𝐶 1⁄2
𝜃𝑜𝑝𝑡 = √𝜃𝑚𝑎𝑥 = (1 + 𝑛 ) (3.15)
0 𝑈𝑟 0 𝑇

𝑆𝑔𝑒𝑛 𝑞0∗ 𝐶 𝜃𝑚𝑎𝑥

=𝜃−2−𝑛 + (3.16)
𝑈𝑟 𝐴𝑟 0 𝑈𝑟 𝑇𝑎𝑝𝑝 𝜃

where 𝐴𝑟 is the receiver area, 𝜃 is dimensionless temperature, 𝑞0∗ is radiation falling on a receiver,
𝑛0 is the collector optical efficiency 𝑈𝑟 is the receiver-ambient heat transfer coefficient. By
combining Eqns. 3.15 with 3.16 the minimum entropy generation rate, Eqn. 3.17 can be
determined, where Kalogirou illustrates (Fig. 3.10) dimensionless entropy generation for three
collector types based on measurements of stagnation temperature: parabolic trough (565 ᵒC),
parabolic dish (1285 ᵒC) and a central receiver (1750 ᵒC).

𝑆𝑔𝑒𝑛,𝑚𝑖𝑛 𝜃 −1
= 2(√𝜃𝑚𝑎𝑥 − 1)𝜃 − 𝑇𝑚𝑎𝑥⁄𝑇 (3.17)
𝑈𝑟 𝐴𝑟 𝑎𝑝𝑝 0

Figure 3.10. Entropy generation & optimum temperatures vs. collector solar concentration ratio
for three solar collector technologies: parabolic trough, parabolic dish and central receiver [55].

47
3.3 Solar Selective Surface Coatings

CSP solar collectors must have high radiation absorptance in the solar energy spectrum, especially
at elevated temperatures above 400 ᵒC, where they may emit a significant amount of thermal
radiation thereby losing heat to the environment [56]. It is also desirable to have long-wave
emittance of the surface as low as possible to reduce losses. At very high temperatures there is an
overlap of the spectral absorptance and emittance, making it difficult to develop a truly selective
absorber. Selective coatings that have high absorptance for solar radiation and high transmittance
for long-wave radiation can be applied to substrates with low emittance. Here the coating absorbs
solar energy and the substrate acts as the poor emitter of long-wave radiation [14]. Additionally,
coatings and paints must endure high temperatures in an oxidizing environment for many thermal
cycles and under potentially large thermal gradients [56].
Generally, surfaces that have a low emittance value often have low absorptance, which also
reduces absorbed solar energy. However, “selective coatings” facilitate surfaces that have a high
absorptance (and emittance) for short wavelength (visible) light and have low average absorptance
and emittance for long wavelength radiation (IR radiation) values during optimal CSP ambient
conditions. These types of surfaces have property capabilities of selective tenability for both α and
ε to optimize heat losses and thermal absorption with solar collectors [56]. As prescribed by the
energy balance in Eqn. 3.18 to maximize the useful heat collected by a collector, high absorptance
and a low emittance values are desired, where based on Kirchoff’s Law, these parameters are only
equal for radiation specific wavelengths:
𝛼𝜆 = 𝜀𝜆 (3.18)

From Eqn. 3.18 the employment of the subscript λ requires that these radiative terms be spectrally
dependent requiring integration over all wavelengths to determine bulk property values.
Principally, selective surfaces do not violate Kirchoff’s law as they have high absorption for short
wavelength radiation and low emittance, particularly in the IR portion of the spectrum, for long
wavelength radiation. The end result is a surface that absorbs solar energy well, but does not radiate
thermal energy very well [56]. This phenomenon is presented in Fig. 3.11 where 99.5% of CSP
solar energy for absorption occurs generally at wavelengths below 5 μm [56]. This idealized
surface is called a semi-gray surface, since it can be considered gray in the solar spectrum.
However, the effect of using selective surfaces on higher temperature absorbers is less dramatic,
where if the collector surface temperature is approximately 700 ᵒC, only 43.6% of its radiated
energy is at wavelengths above 5 μm and the integrated emittance is 64.5%.

Figure 3.11. Radiation properties of a hypothetical selective surface [56].

48
These types of coatings, such as black chrome [60], are generally deposited as thin (2-3 μm) layers,
such as by electro-deposition, where microscopic particles can effectively trap incident light,
making the coating appear black. For black chrome, although it is highly absorbing in the visible
spectrum, it is transparent in the IR region, therefore a collector can take on lower emittance of a
potentially brighter substrate [56]. Additionally, the collector surface area can be modified to
minimize radiative heat losses, where cavity receiver designs can be designed to have small
openings through which concentrated solar energy can pass onto larger absorbing surfaces.
Another heat loss mechanism, particularly with trough collectors is “end-loss” which is the fraction
of energy being reflected from the curved surface that falls beyond the receiver due to the angle of
incidence [56].
Another approach at facilitating solar selective surfaces is through geometrical surface
modifications, as illustrated by Fig. 3.12. Here, directional selectivity can be obtained by proper
arrangement of the surface, where deep grooves that are large relative to all wavelengths of
radiation concerned, can be arranged for radiation from near-normal directions to the overall
surface be reflected within the grooves absorbing a fraction of the beam. This enhanced absorption
surface provides an increase in solar absorptance while increasing the long-wavelength emittance
[9]. Hollands [61] demonstrated how a folded selective surface can have significantly enhance
effective properties by proper configuration, where for a flat –plate collector surface that had α =
0.60 and ε = 0.05, the addition of 55ᵒ grooves, increased the average values to an effective α of
0.90 and an equivalent ε of 0.10.

Figure 3.12. Selective geometrical surface modifications for absorption of radiation by successive
reflections on folded metal sheets [9, 62].

With regards to conduction, this mode of heat loss is generally considered with respect to its cross-
sectional area and temperature gradient as prescribed by Eqn. 3.19. In flat-plate collector designs,
the sides and back surface of an absorber plate can facilitate heat loss, where they should
incorporate good insulation properties (low k and relatively high thickness Δx) to reduce these
losses. However, this heat loss mode is typically insignificant when compared to convective or
radiative losses for any heat collector design, especially at high temperatures.

𝑄̇𝑙𝑜𝑠𝑠,𝑐𝑜𝑛𝑑𝑢𝑐𝑡𝑖𝑜𝑛 = 𝑘̅ ∙ ∆𝑥̅ ∙ 𝐴𝑐𝑟𝑜𝑠𝑠 ∙ (𝑇𝑐 − 𝑇𝑎𝑚𝑏 ) (3.19)

If insulation used within a heat transfer material facilitates a sufficiently high thermal resistance,
the adiabatic approximation can typically be employed, whereby zero heat transfer occurs and Eqn.
3.19 approximately equals zero.

49
3.4 Collector Performance and Efficiency

The optimum azimuth angle for a flat-plate collector, considered to be the simplest form of a solar
collector, is typically 0ᵒ in the northern hemisphere and 180ᵒ in the southern hemisphere [56]. A
detailed overview for the calculation of solar radiation angles for flat-plate collectors can be found
by [9]. There are several common methods for determining flat-plate thermal collector
performance, where calculations and tests are performed on clear days with relatively high and
constant solar irradiance [56]. Empirically, hot heat-transfer fluid is circulated through an absorber
or receiver at a given mass flow rate, 𝑚̇ to raise it to test temperature. For parabolic troughs,
turbulent flow is maintained within the receiver tubes to ensure good heat transfer between the
fluid and the wall of the tubes. A measurement is only made when the collector is at steady state,
which is determined by a constant rise in heat transfer fluid temperature as it passes through the
receiver [48]. There are essentially three methods for assessing collector efficiency for a flat plate
collector, Fig. 3.13: collector balance, system balance and heat loss measurement methods [48].
The collector balance method is considered the most common method used for flat-plate and
parabolic trough collectors where a temperature controller is typically connected to an auxillary
heating and cooling device to ensure a constant fluid inlet temperature can be maintained. The rate
of energy being transferred to the HTF at steady state is provided by Eqn. 3.20.

𝑄̇𝑢𝑠𝑒𝑓𝑢𝑙 = 𝑚̇𝐶𝑝 (𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛 ) (3.20)

Test procedures developed at Sandia National Laboratories for this method are described by
Dudley et al. [63].

a. b.

c.

Figure 3.13. Three experimental methods for measuring collector efficiency based on a. collector
balance, b. system balance and c. heat loss measurement methods [48].

50
For conditions where it is not easy to accurately measure fluid inlet or outlet temperatures or mass
flow rate within the system, an insulated flow loop and tank can be used to facilitate a system
energy balance:
𝑑𝑇
𝑄̇𝑢𝑠𝑒𝑓𝑢𝑙 = 𝑚̇𝐶𝑝 𝑑𝑡 − 𝑊̇𝑝𝑢𝑚𝑝 (3.21)

where dT/dt represents the temperature rate increase of the tank with the HTF and 𝑊̇𝑝𝑢𝑚𝑝
represents the pump work entering the insulated system. For this method the amount of HTF must
be large enough to ensure the bulk temperature rise does not inhibit a steady state temperature
condition [48]. Normally the absorber surface should operate at a higher temperature than the HTF;
however, when heat losses exist, (e.g. a defocused or shaded collector) the absorber surface will
be cooler than the HTF which could result in measurement error that can indicate heat losses that
are lower than would occur during normal operating conditions [48].
A third heat loss measurement method used for CSP collector systems involves two
separate tests to determine thermal performance. Here the rate of optical energy collected by Eqn.
3.8 is measured where the collector HTF is operated close to ambient. Then with the second test,
the collector is operated normally with a heat input where what will be lost from the receiver due
to the collector-ambient temperature difference as described by Fig. 3.13c. Here the test is
performed at varying collector operating temperatures where the rate of useful energy is given as:

𝑄̇𝑢𝑠𝑒𝑓𝑢𝑙 = 𝑚̇𝐶𝑝 (𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛 )| 𝑇𝐻𝑇𝐹−𝑇∞ − 𝑚̇𝐶𝑝 (𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛 )|𝑄̇𝑖𝑛=0 (3.22)

To determine solar collector efficiency, the collector aperture first needs to be aligned as close as
possible to the DNI incident vector to eliminate any uncertainties due to off-normal incidence
angle effects. Additionally, test data are obtained, typically at one temperature to determine the
effect of varying incidence angles on collector performance [56]. Solar energy collection
efficiency of thermal collectors is defined as the ratio of the rate of useful thermal energy leaving
the collector to the useable solar irradiance falling on the aperture area as:

𝑄̇𝑢𝑠𝑒𝑓𝑢𝑙
𝜂𝑐𝑜𝑙 = 𝐼 (3.23)
𝑖𝑛𝑐,𝑎 𝐴𝐴𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒

where the incoming solar irradiance falling on the collector aperture, 𝐼𝑖𝑛𝑐,𝑎 , multiplied by the
collector aperture area, 𝐴𝐴𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 represents the maximum amount of solar energy that could be
captured by that collector. Therefore, incorporation of Eqn. 3.23 into Eqn. 3.20 provides the
thermal collection efficiency for flat plate collectors as:

𝑚̇𝐶𝑝 (𝑇𝑜𝑢𝑡 −𝑇𝑖𝑛 )

𝜂𝑐𝑜𝑙 = 𝐼 (3.24)
𝑖𝑛𝑐,𝑎 𝐴𝐴𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 𝑐𝑜𝑠𝜃𝑖𝑛𝑐

The thermal efficiency for receivers, based on collection efficiency is therefore:

𝛼𝑄̇𝑖𝑛 −𝑄̇𝑜𝑢𝑡 4 +𝑓
𝜀𝜎𝐹𝑣𝑖𝑒𝑤 𝑇𝑟𝑒𝑐 ̅𝑐𝑜𝑛𝑣 (𝑇𝑟𝑒𝑐 −𝑇∞ )
𝑐𝑜𝑛𝑣 ℎ
𝜂𝑡ℎ = =𝛼− (3.25)
𝑄̇𝑖𝑛 𝜂𝑓𝑖𝑒𝑙𝑑 𝐸𝐷𝑁𝐼 𝐶

where a parametric study by Ho and Iverson [64] found that receiver thermal efficiency was more
sensitive to concentration factor C than with the view factor Fig. 3.14.

51
 a. b.

c.

Figure 3.14. Thermal efficiency receiver plots of surface temperature with a. varying
concentration ratio b. radiative view factor and c. convection heat loss [63].

Optical efficiency, as defined in Eqn. 3.26 which is defined as the rate of optical (short wavelength)
energy reaching the absorber or receiver, divided by the appropriate solar resource [56]. It is
determined by measurements while operating the collector at a constant inlet temperature while
on-sun and then with it in a defocused orientation. This same test can be performed on a flat-plate
collector by covering and uncovering the absorber with a sheet of opaque material. The efficiency
of the on-sun collector to the rate of heat loss is divided by the same irradiance and aperture area
with the result being optical efficiency [56].

𝜂𝑜𝑝𝑡 = Γ𝜌𝜏𝛼 (3.26)

Finally, the balance of gains and losses within the receiver provides the total receiver efficiency as
the product of Eqns. 3.25 and 3.26:
𝜂𝑅𝑒𝑐. = 𝜂𝑡ℎ 𝜂𝑜𝑝𝑡 (3.27)

52
Another method for enhancing optical and thermal efficiencies of CSP receivers, without the use
of coatings, is through enhancement of light-trapping using fractal-like geometries. Ho et al. [443]
developed novel fractal-like receiver geometries and features at both macro (meters) and meso
(millimeters to centimeters) scales (Fig. 3.15) to increase light trapping and thermal efficiency at
multiple length scales. According to the authors, advantages of this design approach could include
increased solar absorptance with reduced thermal emittance.

Figure 3.15. Fractal-like receiver geometries at the macro and meso-scales [443].

The results of their analytical/experimental investigation demonstrated that fractal-like structures

and geometries can increase thermal efficiency by several percentage points at both meso- and
macro-scales, depending on factors such as intrinsic absorptance. However, the impact was more
pronounced for materials with lower intrinsic solar absorptance (<0.9) [443]. Additionally,
modelling was also performed for macro-scale receiver geometries, where as illustrated by Fig.
3.16, modelled results found thermal efficiencies increased by nearly 5% with radial or linear
bladed receiver configurations. Radiative losses were reduced with the fractal-like geometries due
to reduced local radiative view factors in the hottest interior regions, while convective losses were
slightly higher in the vertical panel configurations. Additionally, while convective losses were
found to be reduced in the horizontal louvered panel configuration, the authors suggested for this
configuration that heat losses from the higher temperature interior could potentially be recuperated
in cooler regions which could reduce the overall convective heat loss [443].

53
Figure 3.16. Thermal efficiency and heat losses for macro-scale receiver geometries with an
average irradiance of 500 W/m2 [443, 444].

3.5 CSP System Receiver Technologies

Previous work has investigated solar transients and fluctuation in irradiance which can be
mitigated by using oversized mirror fields (with solar multiples higher than 1) and making use of
excess energy to load a thermal or chemical storage system [65]. However, these oversized mirror
fields can increase costs where previous studies have focused on increasing receiver efficiency to
reduce the overall receiver area [66]. Thermal losses for all CSP receivers include primarily
convective and radiative heat losses to the ambient air and conductive heat loss through the
receiver insulation. As previously discussed, radiative losses are dependent on receiver wall
temperatures, shape factors and emissivity/absorptivity of receivers walls while conduction is
dependent on receiver temperature and the type of insulation material employed [65]. Radiant flux
and temperature are substantially higher in solar towers than in parabolic troughs, and therefore,
advanced technology is involved in design as well as high-performance materials that can
withstand high-temperatures. The solar receiver should mimic a blackbody by minimizing
radiation losses where cavities, black-painted tube panels or porous absorbers are able to trap
incident photons. For many designs the solar receiver is a single unit that centralizes all energy
collected by concentrating mirrors [65]. There are different solar receiver classification criteria
depending on construction solution, such as the use of intermediate materials as well as the type
of HTF employed or other heat transfer mechanisms. According to geometrical configuration,
there are essentially two types of design configurations: external and cavity-type receivers. Typical
receiver-absorber operating temperatures are between 500-1200 ᵒC where incident flux levels span
a wide range between 300 to over 1000 kW/m2 [65].
Receivers can also be directly or indirectly irradiated depending on absorber materials used
to transfer energy to the HTF [67]. Directly irradiated receivers make use of fluids or particle
streams able to efficiently absorb the concentrated radiation. Particle receiver designs however
make use of falling curtains or fluidized beds, where darkened liquid fluids can also make use
falling films [65]. The key design element if indirectly heated receivers is the radiative/convective

54
heat exchange surface. Essentially, two heat transfer options are used, tubular panels and
volumetric surfaces. With tubular panels, HTF flows inside the tube and removes heat collected
by the external black panel surface by convection, therefore acting as a recuperative heat
exchanger. Based on the HTF properties and incident solar radiation, the tube may also undergo
thermomechanical stress, and since heat transfer is through the tube surface, it may also be difficult
to operate at an incident flux above 600 kW/m2 [65]. As shown in Table 3.3 by Prakash et al. [65]
high thermal conductivity liquids are able to better achieve operating flux levels above 1 MW/m2
when compared to other common CSP working fluids.

Table 3.3. Operating temperature and flux ranges of solar tower receivers [65].
Liquid
Fluid Water/Steam Molten Salt Volumetric Air
Sodium
Average Flux (MW/m2) 0.1-0.3 0.4-0.5 0.4-0.5 0.5-0.6
Peak Flux (MW/m2) 0.4-0.6 1.4-2.5 0.7-0.8 0.8-1.0
Fluid Outlet Temp. (ᵒC) 490-525 540 540-565 700-800

The operating temperature, heat storage system and thermodynamic cycle influence the design and
can make selection of a receiver technology very complex. In general, according to Prakash et al.
[65] tubular technologies allow either high temperatures (up to 1000 ᵒC) or high pressures (up to
120 bar), but not both [68]. Directly irradiated or volumetric receivers allow even higher
temperatures but limit pressures to below 15 bar [65].

3.5.1 Direct Absorption Receivers

3.5.1.1 Cavity Receivers

Cavity receivers are widely used especially with solar dish systems for providing industrial heat
[69, 70] for generating electric power [71, 72] and for thermochemical reactions [73]. The cavity
receivers used with these systems for process heat applications are generally cylindrical in shape,
operate at medium temperatures (< 300 ᵒC), have 𝑑/𝐷 ratios even greater than one, and may have
non-uniform wall temperatures [77]. When compared to other types of receivers, cavity type
receivers however have been shown to significantly reduce thermal and optical losses [74]. This
receiver, as shown in Fig. 3.17a, typically consists of a receiver tube assembly enclosed within a
cylindrical open cavity. Typically, the plane of the cavity aperture is placed near the focus of the
parabola and normal to the axis of the parabola (also referred to as a focal plane receiver) Fig.
3.17b. Although the cavity could be linear and used with parabolic troughs, for parabolic dishes,
a cavity receiver does not have stringent constraints on tube volume and pressure drop. In a cavity
receiver, radiation reflected from heliostats pass through an aperture into an enclosed structure
before impinging on the heat transfer surface. Cavities are constrained angularly and external
receivers can be designed to be flat-plate tubular or a cylindrically shaped.

55
 b.

a.

Figure 3.17. a. General overall solar cavity receiver design [75] and b. side profile of receiver
cavity [76].

The effective absorptance of a cavity without a cover on its aperture, where the fraction of incident
radiation absorbed by the cavity, is a function of absorptance of the inner surfaces and the ratio of
the areas of the cavity as prescribed by Eqn. 3.28:
𝛼
𝛼𝑒𝑓𝑓 = 𝛼 +(1−𝛼𝑖 )𝐴 (3.28)
𝑖 𝑖 𝑎 ⁄𝐴𝑖

where 𝛼𝑖 is the absorptance of the inner surface of the cavity, Aa is the area of the aperture and Ai
is the area of the inner surface. Duffie and Beckman recommends as an approximation that 𝛼𝑖 can
be evaluated at the effective angle of incidence of diffuse radiation, around 60ᵒ [9].
The overall efficiency of these systems is dependent on the thermal losses occurring from
the cavity receiver, where it is observed from literature that convective heat transfer constitutes the
largest share of thermal losses [77-79]. These losses are found to be dependent on receiver
inclination θinc,rec receiver wall boundary condition, aspect ratio (L/D), opening ratio (d/D) and
magnitude/direction of the external wind [77]. Experimental and numerical investigations on
natural convection losses in cavity receivers used with solar dish concentrators have been
performed on different cavity shapes. Stine and McDonald [9] have performed experimental
studies on conical frustum-cylindrical cavity receivers while Kugath et al. [445] has performed
field studies on a similar receiver shape at medium temperature levels. Cylindrical shaped cavity
receivers have been analyzed experimentally and numerically [65, 78, 80]. Numerical
investigations on hemispherical cavity receivers have been carried out by Sendhil, Kumar and
Reddy [79, 81]. Prakash investigated the impacts of wall temperatures, receiver inclination, and
𝑑/𝐷 ratio on natural convection losses and developed Nusselt number correlations to assess losses
due to convection [77].
Radiation and conduction heat losses from the receiver can be predicted reasonably well
by analytical techniques where convection from the cavity is much more complicated and
generally not amenable to analytical predictions [76] due to wind effects and varying receiver
orientation. Therefore, convective heat loss from cavity receivers is typically determined
empirically [65]. Design correlations for estimating natural convective heat loss from cavity

56
receivers are usually derived experimentally. For cavity receivers convective heat loss can be
directly calculated as [82]:
𝑄̇𝐶𝑜𝑛𝑣 = 𝑄̇𝐹𝑜𝑟𝑐𝑒𝑑 + 𝑄̇𝑁𝑎𝑡
𝐴
𝑄̇𝐹𝑜𝑟𝑐𝑒𝑑 = 7631 𝑊 0.2 (3.29)
𝑎𝑝

𝑄̇𝑁𝑎𝑡 = 5077𝐴𝐶𝑎𝑣

where A is the aperture area, Wap is the aperture width, and Acav is the total area inside the cavity.
Further detailed convection loss correlations for other types of solar receivers can be found by
Becker and Vant-Hull [67]. Koenig and Marvin [83] performed experiments where they developed
correlation Eqns. 3.30 and 3.31 for average natural convection. Stine and McDonald [80] found
however that natural convective heat loss is better predicted if the coefficient of 0.78 is used instead
of 0.52 and if the full geometric surface area of the cavity is used.
̅
̅̅̅̅𝐿 = ℎ𝐿 = 0.78𝑃(𝜃)𝑙𝐶1.75 (𝐺𝑟𝐿 𝑃𝑟)0.25
𝑁𝑢 (3.30)
𝑘𝑓

𝑄̇𝐶𝑜𝑛𝑣. = ℎ̅𝐿 𝐴(𝑇𝐶𝑎𝑣𝑖𝑡𝑦 − 𝑇∞ ) (3.31)

where all fluid properties are evaluated at T = 11/16TCavity+T∞, 𝜃 is the receiver tilt angle, 𝛽 is the
coefficient of thermal expansion of air equal to 1/T, where 𝑙𝐶 = 𝑅𝑎𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 ⁄𝑅𝐶𝑎𝑣𝑖𝑡𝑦 , 𝐿 =
√2𝑅𝐶𝑎𝑣𝑖𝑡𝑦 and TCavity is the average temperature of the cavity heat transfer tubing. Additionally,
the following parameters are defined as follows:

𝑃(𝜃) = 𝑐𝑜𝑠 3.2 𝜃 𝑓𝑜𝑟 0° ≤ 𝜃 ≤ 45°

(3.32)
𝑃(𝜃) = 0.707𝑐𝑜𝑠 2.2 𝜃 𝑓𝑜𝑟 0° ≤ 𝜃 ≤ 45°

Siebers and Kraabel [84] reported the following correlation for predicting turbulent natural
convection from a central receiver cavity, over a range of 105≤GrL≤1012:

0.18
̅̅̅̅𝐿 = 0.088𝐺𝑟𝐿1⁄3 ( 𝑇𝑤 )
𝑁𝑢 (3.33)
𝑇 ∞

where Tw is the average internal wall temperature and L is the characteristic height of the cavity.
Eqn. 3.33 was derived based on results of a large 2.2 m cubical cavity experiment developed by
Kraabel [85], and experiments of 0.2 m and 0.6 m cubical cavities performed by LeQuere, Penot
and Mirenayat [86]. Stine and McDonald [80] later performed natural convective cavity receiver
heat loss experiments that included effects of different receiver temperatures, tilt angles and
aperture sizes, which provided the correlation:
0.18 𝑑 𝑠
̅̅̅̅𝐿 = 0.088𝐺𝑟𝐿1⁄3 ( 𝑇𝑤 )
𝑁𝑢 (𝑐𝑜𝑠𝜃)2.47 ( ) (3.34)
𝑇 ∞ 𝐿

where 𝑠 = 1.12 − 0.982(𝑑 ⁄𝐿), d is the aperture diameter and L is the receiver internal diameter
at the cylindrical region. For this correlation the heat transfer area depends on whether solar

57
insolation is present, however for off-sun applications, only the portion of the receiver interior
geometric surface area covered with heat transfer tubing should be considered [76].
Siebers and Kraabel [84] suggest as a first approximation that forced and natural
convection components from a cavity receiver be simply added together to obtain the total
convective heat loss. Clausing [87] performed simplified numerical experiments which calculated
convective heat losses in a large central cavity receiver based on first law principles and the energy
transferred from the hot receiver interior walls to the air inside the cavity. They also considered
the energy transfer across the aperture by the combined influences of flow over the aperture due
to wind and buoyancy-induced flow due to the cooler ambient air. The results of the numerical
work by Ma [76] suggest that the influence of wind at 18 mph or less is minimal. This finding is
in agreement with the experimental results of McMordie [88] who examined wind effects on
convection from central cavity receivers. As a base line analysis, Fig. 3.18, Ma analyzed 3 different
wind speeds on a receiver with a temperature of 277 ᵒC and found that convective heat loss from
the receiver was highest with the receiver facing horizontally (0°) and lowest with the receiver
facing straight down (90ᵒ). With the receiver in a horizontally-facing orientation, natural
convective heat loss was approximately 2 kW and with the receiver facing straight down these
losses were presumed to be zero [76].

Figure 3.18. Natural convection heat loss from a cavity receiver at 277 ᵒC [76].

The natural convective heat loss contribution was found to decrease as the receiver was tilted
downward due to a larger portion of the receiver volume being in the stagnant zone as illustrated
in Fig. 3.19. The authors found that convective currents are virtually non-existent and air
temperature is high, as well as a smaller portion being in the convective zone where significant air
currents exist. It has been observed by Siebers and Kraabel [84] and Clausing [87] that the interior

58
volume above the horizontal plane passing through the uppermost portion of the aperture is
relatively stagnant, consisting of high-temperature air.

Figure 3.19. Cavity receiver diagram of stagnant and convective zones [76].

During nominal operation, natural convective currents flow inside the receiver from bottom to top,
in a vertical plane. Several investigators [65, 76] have found that the orientation of a cavity receiver
with respect to wind magnitude and direction can have a significant impact on forced convection
heat losses. Numerical investigations by Prakash et al. [65] examined convective heat losses from
a downward facing cylindrical cavity receiver, over varying inlet fluid temperatures between 50-
300 ᵒC and receiver inclinations between 0-90°. The investigators found that the convective losses
increase with mean receiver temperature and decrease with inclination. The authors also found that
wind induced convective losses are generally higher than with no wind (varying between 22-75%
for 1 m/s wind speed and between 30-140% for 3 m/s wind speed). Fig. 3.20 provides a parametric
view of convection phenomena for different wind directions and cavity tilt angles.

a. b.

c. d.

Figure 3.20. Graphical depiction of natural convection zones within a cavity receiver for tilt angle
orientations of a. 0ᵒ, b. 30ᵒ, 60ᵒ and 90ᵒ [76].

59
For side-facing winds, forced convective currents are generally in a direction normal to the plane
of natural convective currents. Because of this orthogonal relationship between natural and forced
convective currents, it is reasonable to approximate that forced convection from the receiver is
independent of natural convection. In addition, pure forced convection should change minimally
as the receiver tilt angle changes. In the absence of gravity, side-on wind convective heat loss
would be the same for any receiver tilt angle [76]. From data collected for side-on winds, Ma
produced the Eqn. 3.35 curve fit for pure forced convection as a function of wind speed:

ℎ̅𝑓𝑜𝑟𝑐𝑒𝑑 = 0.1967𝑉 1.849 (3.35)

where V is the velocity of side-on directional wind. This equation is based on full interior geometric
surface area of the receiver, where the exponential value of 1.849 is much larger than the value of
0.8 provided by the general Nusselt correlation for turbulent heat transfer from a flat plate:
̅
̅̅̅̅𝐿 = ℎ𝐿 = 0.037𝑅𝑒𝐿0.8 𝑃𝑟 1⁄3
𝑁𝑢 (3.36)
𝑘 𝑓

According to Ma [76] the value of 1.849 is closer to that normally associated with shear stress,
which suggests that the determining factor for heat transfer from the cavity may be the ability of
wind to transfer mass and energy across the aperture via fluid shear and not necessarily the ability
of the receiver walls to transfer energy to the air inside the cavity, which is consistent with findings
by Clausing [87]. For head-on winds with the assumption that both natural and forced convection
components are additive, Ma found that increased convective head loss was described by:

ℎ̅𝑓𝑜𝑟𝑐𝑒𝑑 = 𝑓(𝜉)𝑉 1.401 (3.37)

𝑓(𝜉) = 0.1634 + 0.7498𝑠𝑖𝑛𝜃 − 0.5026𝑠𝑖𝑛2𝜃 + 0.3278𝑠𝑖𝑛3𝜃 (3.38)

McMordie found that for winds of 3 to 15 mph, wind speed and direction effects were also
indistinguishable. However, Kugath et al. [69] measured effects of a 10 mph wind on convective
heat loss from a cavity receiver and found convective heat loss to be highly dependent upon
receiver orientation. The highest convective heat loss for that study was observed with wind
blowing directly into the cavity, being as much as four times the level of natural convection. They
also found that for wind blowing directly behind the receiver, total convective heat loss was not
much higher than pure natural convection. An experimental investigation conducted by Faust et
al. [89] showed that a noticeable increase in receiver convection occurred with a wind speed of
only 2 mph. In their experiment the investigators observed winds parallel to the aperture plane
result in the highest convective heat loss, where the authors concluded that with wind blowing in
this direction, the aperture lies in the separation region and is subjected to the suction pressure of
the air flow. However, the authors also found that winds perpendicular to the aperture plane were
convective heat loss because flow stagnation supposedly decreases the responsible for natural
convection.
Solar cavity receivers used for high temperature applications are generally modeled as
plain walls with isothermal boundary conditions and have 𝑑/𝐷 ratios equal to or less than one [79,
90]. Numerical modeling of the cylindrical cavity receivers encountered in process heat
applications (< 300 ᵒC) is limited. Numerical studies performed by Prakash et al. [65, 91] have
been carried out on such a cavity receiver. The use of Fluent CFD software for modeling the

60
receiver tubes and the flow within the tubes has been demonstrated by these studies for the cavity
receiver having 𝑑/𝐷 ratio greater than one. It is therefore noticed that numerical study of natural
convection loss from cylindrical cavity receivers of various diameters used for process heat
applications at temperature levels of about 150ᵒC–250ᵒC having 𝑑/𝐷 ratio of 1 and 0.5 and inclined
at different angles has not been performed. The use of Fluent CFD software for modeling the
receiver tubes and the flow within the tubes has been demonstrated by these studies for the cavity
receiver having 𝑑/𝐷 ratio greater than one.

3.5.1.2 Direct Absorption Particle Receivers

In a solar receiver, heat is transferred to an HTF, which transports this energy to a conversion sub-
system that includes heat storage, heat exchangers, and a power block. Since existing CSP HTFs
have drawbacks due to limited working temperatures such as salts (~240-565 ᵒC), corrosion and
safety limitations such as liquid metals, solid particles as an HTF have been demonstrated to reach
operating temperatures as high as 750 ᵒC and flux densities as high as 250 kW/m2 [92]. Preliminary
studies have shown that for a receiver design power of 395 MWth and an outlet temperature of 800
°C a receiver efficiency of approximately 90 % can be achieved. Additionally, by applying a
favorable operation strategy, a receiver efficiency of up to 67 % at 20%-part load condition can be
possible [93]. The optimum applicable power range for one receiver is expected to be between 50
and 400 MWth [94]. The maximum receiver temperature can be found in the HTF and on the
receiver material which leads to decreased radiation losses and material stress and fatigue. Open
particle receivers currently feature solid particles surrounded by a gaseous fluid. This can have
several advantages such as increased radiation absorption, high heat capacity or lower material
temperature towards the surroundings. Different particulate materials and designs have been
studied for these systems, including direct-heating receivers with free-falling particles, Fig. 3.21a
or obstructed particle flows Fig. 3.21b.

a. b.
Figure 3.21. a. Illustration of a direct solid-particle receiver with hydrodynamic and thermal
processes [96] and b. obstructed particle receiver flow design [95].

61
In another embodiment or the falling particle receiver Wu et al. [94] considered a cavity particle
receiver with a non-rotating feed hopper inlet Fig. 3.20a and a centrifugal particle receiver Fig.
3.21b with rotating particle inlet. In their study the authors use bauxite particles as an HTM that
had an absorption coefficient of approximately 0.8-0.9 and a heat capacity between 1000-1250
J/kg-K for the investigated temperature range. As shown in Fig. 3.20a the particles were
transported to a feed hopper at the top of a face-down cylindrical cavity. While the particles fall
down through an inlet slit to a collecting ring at the receiver bottom, a free falling curtain parallel
to the inner cavity wall is formed. Solar radiation enters the receiver through the open aperture and
is directly absorbed by the particle curtain. Similar to other cavity receivers, this design took
advantage of reduced radiation losses due to the face-down design while allowing the particle
curtain to be protected against the impacts of wind. An efficient recirculation strategy was also
leveraged to realize optically dense curtains for high absorptivity and hence to increase the receiver
efficiency. From their Matlab-based preliminary CFD models the instigators found for a face-down
cavity, with a design power of 3 MWth, that wind speeds between 5–15 m/s with particles lead to
convective losses around 2.5–6.5 %. These losses were found to be significant, where
consideration is required especially for large-scale power plants with tower heights up to 300 m
[94].
A second particle receiver concept, Fig. 3.22b, entails a centrifugal particle receiver that
combines high receiver efficiency which could be well-suited for decentralized small-scale CSP
applications in the power range of 0.1 to 1 MWth, as well as for systems up to 200 MWth.
Preliminary analysis for a 1 MWth plant found for diameter to height ratio of 1.5, reflection losses
of approximately 3% were determined. Using the Stefan-Boltzmann law the expected radiation
losses for an aperture size of 1.13 m2, over a temperature range between 600-1000 ᵒC, were 9%
[94]. Additionally, the investigators suggested that receiver rotation was able to reduce convection
losses.

a. b.
Figure 3.22. Direct particle receivers by DLR for a. non-rotating particle injection cavity receiver
and a b. rotating inlet centrifugal particle receiver. [94].

With regard to demonstrated falling particle receiver performance, at the National Solar Thermal
Test Facility (NSTTF), Sandia National Laboratories recently completed a 3½-year project funded
by the DOE SunShot Initiative to develop a high-temperature falling particle receiver, Fig. 3.23.
Here, sand-like ceramic particles are heated as they fall through a beam of highly concentrated

62
sunlight [97]. The falling-particle receiver enables concentrated solar power with thermal storage
for on-demand electricity production and process heat at significantly high temperatures (up to
1000 °C and higher), which can increase power cycle efficiencies and reduce levelized costs.
Sandia constructed and successfully demonstrated this continuously, recirculating high-
temperature 1 MWth falling particle receiver achieving peak particle temperatures over 900°C with
bulk temperatures over 800 °C [97]. The particle heating rate reached 100–300°C per meter of
illuminated drop distance, at concentrated solar irradiances of approximately 1,000 kW/m2 and
thermal conversion efficiencies of approximately 80% [97].

Figure 3.23. Falling particle receiver at Sandia National Laboratories [97].

A design approach for CSP particle receivers is through fluidization of falling gas-particle
mixtures. Fluidized flow heat transfer is the object of much research, especially in the area of
combustion due to the potential of high heat transfer to immersed surfaces, which can reach
temperatures on the order of 1000 K [98]. For large thermal input conditions, large gas flows are
required when the temperature is raised to a point where agglomeration occurs [99, 100]. However,
as agglomeration intensifies, convection heat transfer coefficients have been shown to reduce [100]
due to decreased particulate mixing. This presents a fundamental challenge for particles at lower
flow velocities which are required to enhance residence time. For fluidization, there are four
primary modes of heat transfer [101]: bed-to-surface, inter-particle, gas-to-particle, and particle-
to-particle heat transfer. Many studies have investigated bed-to-surface heat transfer, such as the
work by Basu and Nag [102] who divided the overall heat transfer coefficient into components of
particle conduction and gas convection. Particle convection has been previously modeled by a film
model by Heered [103] and Mahalingam and Kolar [104], a packet or cluster renewal model by
Mickley and Fairbanks [105] and Subbaro and Basu [106], single particle model by Brotteril and
Williams [107] and Martin [108]. Gas-to-particle heat transfer components have been modeled by
Eckert and Drake [109] and Wen and Miller [110]. Additionally, particle-to-particle heat transfer
can result from three main mechanisms: heat transfer by radiation, heat conduction through the
contact points between the particles, as well as heat exchange through the gas layer separating the
particles. The first mechanism is only significant at temperatures higher than 600 ᵒC, while the
second occurs when moving particles collide and heat is conducted through the impact area.

63
Gas-solid particle systems in CSP applications use fluidized-bed technology for heat exchange and
packed particles for thermal energy storage (TES) applications. When particle-TES is
implemented in a fluidized-bed (FB) CSP plant, hot-solid particles discharge from the TES hot
storage system to a FB heat exchanger for power generation. The separated outlet cold particles
from the power block system then recirculate through a particle receiver for solar heat absorption
and for charging the particle-TES [111].
Analytical modeling has increasingly been accepted as an effective method to study heat
transfer phenomena of particle-fluid systems [112]. Current approaches for modelling particle flow
and thermal behavior can be classified into two categories: the continuum approach (Eularian-
Eularian) at a macroscopic level, and a discrete modelling approach at a microscopic/particle level
(Eularian-Lagrangian). Table 3.4 provides a recommended modelling approach for various
gas/solid models and scales.

Table 3.4. Model recommendations of solid particle and gas phase dynamics [113].

For the continuum approach, the macroscopic behavior is described by conservation equations of
mass, momentum and energy that are then closed with constitutive relations with initial and
boundary conditions [114-116]. One of the most widely used models, the two-fluid model (TFM),
accounts for both fluid and solid-phases, treated as interpenetrating continuum media in a
computational cell, which is much larger than individual particles but still small compared with
the size of process equipment [114]. However, it’s utility depends on the constitutive or closure
relations for the solid phase and the momentum exchange between phases which is often difficult
to obtain [112].
The discrete approach however considers a finite number of discrete particles interacting
by means of contact and non-contact forces. When coupling with fluid flow, DPS is often coupled
within a computational fluid dynamics (CFD) framework [112]. Here, the flow of a continuum
fluid can be determined by solving Navier-Stokes equations while coupling CFD and DPS through
particle-fluid interaction forces [117, 118]. The main advantage of DPS-CFD is that it can generate
detailed particle-scale information, such as the trajectories and forces pertaining to individual
particles, important for determining mechanisms governing complex flow behavior [112]. With
the DPS-CFD approach, information such as particle-particle or particle-wall contact, local
porosity and local gas-solid flow structure can be produced, which is essential in determining heat
transfer behavior of individual particles [112]. Chen et al. [96] developed a CFD model for an
open particle receiver in the form of a curtain of ceramic particles between 200 μm and 600 μm in
size, where their simulations show good agreement with experimental results. The calculated
receiver efficiencies were found to be approximately 70% for particle outlet temperatures of less
than 1000 K. Cocker and Miller [119] modeled a cylindrical volumetric receiver with air and
carbon nanoparticles as the HTF. Their initial CFD simulations suggest fluid outlet temperatures

64
of up to 1430 K. As new particle and receiver material stability issues are resolved for temperatures
greater than 750 ᵒC [95], thermal efficiencies of solar particle receivers are expected to reach 90%
[120], however there are still heat transfer issues, such as particle-fluid and particle-wall-fluid
interactions that still need to be resolved.
Zhou et al. [121, 122] proposed a comprehensive model taking into account most known heat
transfer mechanisms. This approach considers particle-fluid convection, particle-particle
conduction, and particle radiation. The extended DPS-CFD model offers a useful numerical
technique to elucidate the fundamentals governing heat transfer in packed/fluidized beds at a
particle scale. Overall, various methods have been developed for DPS simulation [123]. In
general, a particle passing through a particle-fluid flow system can either exhibit translational
and/or rotational motion, determined by Newton's second law of motion [112]. The corresponding
governing equations for a particle j acting on a particle i that has a radius r, mass m and moment
of inertia I can be written as:

⃑𝑖
𝑑𝑉 𝑝 𝑁
𝑚 = 𝑓𝑖𝑛𝑡,𝑖 + ∑𝑗=1(𝑓𝑒,𝑖𝑗 + 𝑓𝑑,𝑖𝑗 ) + 𝑚𝑖 𝑔 (3.39)
𝑑𝑡

𝑑𝜔𝑖 𝑁
𝑝
𝐼𝑖 = ∑𝑗=1(𝑀𝑡,𝑖𝑗 + 𝑀𝑟,𝑖𝑗 ) (3.40)
𝑑𝑡

where 𝑉 ⃑ and ω are the translational and angular velocities of the particle respectively. The forces
𝑓 considered within these equations are: particle-fluid interaction force 𝑓𝑖𝑛𝑡,𝑖 , the gravitational
force 𝑚𝑖 𝑔, and inter-particle forces between particles which include the elastic force 𝑓𝑒,𝑖𝑗 and
viscous damping force 𝑓𝑑,𝑖𝑗 . These inter-particle forces can be resolved into the normal and
tangential components at a contact point [123]. The torque acting on particle i by particle j
includes two components: 𝑀𝑡,𝑖𝑗 which is generated by tangential force and causes particle i to
rotate, and 𝑀𝑟,𝑖𝑗 which is generated by asymmetric normal forces that slows the relative
rotation between particles. A particle may undergo multiple interactions where the individual
interaction forces and torques are summed over the Np particles interacting with particle I [123]
Interaction forces and torque determined between two particles (approximated as spheres)
have previously been established by several investigators [123]. According to Zhou et al. [112]
there are approximately nine particle-particle interaction forces which impact solid-particle heat
transport and are non-linear. Additionally, many correlations are available in the literature to
calculate the fluid drag acting on individual particles [124, 125]. The heat transfer of particle i and
its surroundings is considered to be controlled by three mechanisms: particle-fluid convection,
particle-particle or particle-wall conduction, and radiation. According to the energy balance, the
governing equation for particle i can be generally written as [112]:

𝑑𝑇𝑖 𝑁𝑝
𝑚𝑖 𝐶𝑝,𝑖 = ∑𝑗=1 𝑄̇𝑖,𝑗 + 𝑄̇𝑖,𝑓 + 𝑄̇𝑖,𝑟𝑎𝑑 + 𝑄̇𝑖,𝑤𝑎𝑙𝑙 (3.41)
𝑑𝑡

where Np is the number of particles exchanging heat with particle 𝑄̇𝑖,𝑗 is the heat exchange rate
between particles i and j due to conduction. Additionally, 𝑄̇𝑖,𝑓 is the heat exchange rate
between particle i and its local surrounding fluid, 𝑄̇𝑖,𝑟𝑎𝑑 is the heat exchange rate between
particle i and its surrounding environment by radiation, and 𝑄̇𝑖 ,wall is particle-wall heat exchange
rate. Cp,i is the particle specific heat.

65
Convective heat transfer between particles has been previously studied with various equations
proposed [126, 127] where originally the convective heat transfer rate between particle i and fluid
f was calculated according to Eqn. 3.42.

𝑄̇𝑖,𝑓 = ℎ𝑖,𝑐𝑜𝑛𝑣 𝐴𝑖 (𝑇𝑓,𝑖 − 𝑇𝑖 ) (3.42)

where 𝐴𝑖 is the particle surface area, 𝑇𝑓,𝑖 is fluid temperature in a computational cell where
particle i is located, and ℎ𝑖,𝑐𝑜𝑛𝑣 is the convective heat transfer coefficient. ℎ𝑖,𝑐𝑜𝑛𝑣 is associated
with the Nusselt number for air, which is a function of particle Reynolds number and gas Prandtl
number:
ℎ 𝐷𝑖
𝑁𝑢𝑖 = 𝑖,𝑐𝑜𝑛𝑣 = 2.0 + 𝑎𝑅𝑒𝑖𝑏 𝑃𝑟 1⁄3 (3.43)
𝑘𝑓

where 𝑘𝑓 and 𝐷𝑖 are the fluid thermal conductivity and particle diameter, respectively. Rei is the
local relative Reynolds number for a particle i. The constant, 2.0, represents the contribution by
particle-fluid natural convection where a and b are parameters determined empirically [112].
Kunii and Levenspiel [127] suggest a value of b = 0.5, with a potential range based on bed
conditions of 0.6-1.8. For heat transfer at the fluid-wall interface Eqn. 3.44 is suggested [112] for
determining the heat transfer coefficient hf,wall where D is the hydraulic diameter, and the
exponent n is 0.4 for heating, and 0.3 for cooling [128].

ℎ𝑓,𝑤𝑎𝑙𝑙 𝐷
𝑁𝑢𝑖 = = 0.023𝑅𝑒 0.8 𝑃𝑟 𝑛 (3.44)
𝑘𝑓

For conduction heat transfer between particles, Fig. 3.24, there are two primary mechanisms that
are generally considered by several authors [129, 130] which include: a. particle-fluid-particle
conduction and b. particle-particle conduction as described by Zhou et al. [121].

Figure 3.24. Diagram for the relative positions of two particle spheres that are in a. non-contact
and b. contact with overlap [121].

A model developed by Cheng et al [130] and Zhou et al. [112] is provided by Eqn. 3.45 where the
heat flux between two particles i and j are approximated as spheres and where 𝑘𝑝𝑖 and 𝑘𝑝𝑗 are the
respective thermal conductivities.

66
𝑄̇𝑖,𝑗 = (3.45)
𝑟𝑠𝑓
2𝜋 ∙ 𝑟𝑑𝑟
(𝑇𝑗 − 𝑇𝑖 ) ∫
𝑟𝑠𝑖𝑗 (√𝑅 2 − 𝑟 2 − 𝑟 (𝑅 + 𝐻)⁄𝑟𝑖𝑗 ) ∙ (1⁄𝑘𝑝𝑖 + 1⁄𝑘𝑝𝑗 ) + 2 [(𝑅 + 𝐻) − √𝑅 2 − 𝑟 2 ]⁄𝑘𝑓

For this formulation conduction heat transfer between particles occurs through the contact area or
between particles and the wall, due to elastic deformation based on two mechanisms: 1. conduction
due to particle-particle static contact (packed bed) and 2. conduction due to particle-particle
collision (typically associated with a fluidized bed) [112]. For static particle-particle conduction a
formulation by Cheng et al. [130] can be employed:

4𝑟𝑐 (𝑇𝑗 −𝑇𝑖 )

𝑄̇𝑖,𝑗 = (1⁄𝑘 +1 ⁄𝑘
(3.46)
𝑝𝑖 𝑝𝑗 )

Conversely, for particle-particle heat transfer due to collision, a modified formulation of the Sun
and Chen [131] model by Zhou et al. [112] can be used:
−1⁄2
(𝑇𝑗 −𝑇𝑖 )𝜋𝑟𝑐2 𝑡𝑐
𝑄̇𝑖,𝑗 = 𝑐′ −1⁄2 −1⁄2 (3.47)
(𝜌𝑝𝑖 𝑐𝑝𝑖 𝑘𝑝𝑖 ) +(𝜌𝑝𝑗 𝑐𝑝𝑗 𝑘𝑝𝑗 )

where rc and tc are particle-particle contact radius and contact duration respectively. As prescribed
by Zhou et al. [112] for a particle-wall static or collision contact, a wall can be treated as a particle
with an infinite diameter and mass.
With regard to radiation heat transfer contributions, an environmental temperature is
generally assumed to represent the enclosed surface temperature around a particle according to
Eqn. 2.21 where the parameter T can be taken as an averaged temperature between the particles
and the fluid by volume fraction in an enclosed spherical domain Ω provided by [121] where 𝑇𝑓,Ω
and 𝑘Ω are the fluid temperature and number of particles located in the domain Ω.

𝑘Ω 1
𝑇 = 𝑇𝑙𝑜𝑐𝑎𝑙,𝑖 = 𝜀𝑓 𝑇𝑓,Ω + (1 − 𝜀𝑓 ) 𝑘 ∑𝑗=1 𝑇𝑗 (𝑗 ≠ 𝑖) (3.48)
Ω

For high temperature gas-particle systems, radiation heat transfer can become the dominant mode
of energy transport, where it is important to resolve the processes of absorption, emission, and
scattering for an investigated media. However, particle-gas prediction modelling may also take a
stochastically distributed [448, 449] approach, where attenuation of radiation in a randomly-
distributed particulate medium can be derived using the Poison Eqn. 3.49:

𝜆𝑁 𝑒𝑥𝑝(−𝜆)
𝑃𝑁 = (3.49)
𝑁!

where PN is the probability, and 𝜆 is the expectation of the number of particles, N in a specified
volume. However, the attenuation of radiation in a particulate medium can deviate from
traditionally-employed Poison distributions, as illustrated by Fig. 3.25a, due to impacts related to
time evolution and/or external forcing such as with turbulent flow [450, 451]. During turbulent
particle-laden flows, particles are carried by the surrounding fluid that may undergo frequent
velocity fluctuations. This can impact the particle inertia and cause particle velocities to lag the

67
local fluid velocity, leading to a segregation of particles out of regions of high fluid vorticity and
into regions of high shear [447].

a. b.
Figure 3.25. Illustrative comparison of an a. Poison particle distribution and b. turbulence-
clustered particle distribution [447].

Radiation transport can be further complicated as the participating medium is not continuous, as
in the case of a cloud of small particles [447] or clusters, Fig. 3.25b. Resolving the interaction
between the radiation and the individual particles in a very large system is impractical, whereas
continuum-based representations of the particle field lend themselves to efficient numerical
techniques based on the solution of the radiative transfer equation. In an investigation by Frankel
et al. [447] the investigators assessed radiation transport through discrete and continuum-based
representations of a particulate medium. Exact solutions for radiation extinction were developed
using a Monte Carlo model in different particle distributions which were then projected on to a
concentration field of varying sizes. The Bouguer–Beer law was also applied to the model to
compute transmission through the domain using a concentration field, determined from the same
particle fields. A solution for radiative transport was validated with a particle-resolved Monte
Carlo experiment where computed transmission was shown to produce deviations from the
Bouguer–Beer law based on the assumption of a uniform particle concentration [447].
Fluidized bed heat transfer at a microscopic scale has also been examined experimentally
[132-134]. In these studies, the authors described heat transfer to a particle by at least three
mechanisms: convection from fluid, conduction from particles or wall, and radiation where the
heat transfer of a particle was found to be strongly affected by the local gas-solid flow structure,
and therefore vary spatially and temporally. This suggests that information derived for a single
particle may not be reliable due to difficulty in quantifying such local structures in a particle bed.

3.5.1.3 Advanced Magnetic Falling Particle Receivers

Although previous falling particle receiver designs have proven outlet temperatures above 800 ᵒC,
and thermal efficiencies between 80-90%, performance challenges still exist to operate at higher
concentration ratios above 1000 suns and greater solar absorptance levels. To increase
absorptance, these receivers will require enhanced particle residence time within a concentrated
beam of sunlight. However, depending on particle size and external forces (e.g., external wind and
flow due to convective heat losses), optimized particle flow can be severely affected, which can
reduce receiver efficiency. To reduce particle flow destabilization and increase particle residence

68
time on the receiver an imposed magnetic field can be utilized based on a collimated design for
three different methodologies [135].
Generally, magnetohydrodynamics (MHD) considers fluid flow of electrically conducting
media in the presence of an electro-magnetic field where previous CSP research has included
electrically conducting liquid metal fluid flow in the presence of a magnetic field [136]. However,
many of these studies [137] considered fluidized solid particles operating temperatures under 600
ᵒC. To date, research does not exist for CSP applications that leverage magnetic falling particles
as the HTM. However, previous work [138] has shown that increasing heat-flux residence time
and inter-particle collisions for particles, with a high heat capacitance, can increase heat transfer
absorption rates considerably. Increases in particle mass loading ratios (seeded particle specific
mass to carrying gas specific mass) have also previously found heat exchanger enhancement by
orders as high [138] as 10, which could have implications for CSP falling particle receivers.
There are several classes of magnetic particles that can be leveraged for CSP falling particle
applications. The main distinction between these types is that some have no collective interaction
of atomic magnetic moments, whereas in other materials there is a very strong interaction between
atomic moments [135]. Ferromagnetic particles exhibit long-range magnetic order below a certain
critical temperature. Even though electronic exchange forces in ferromagnets are very large,
thermal energy eventually overcomes the exchange and produces a randomizing effect. This occurs
at a particular thermal limit called the Curie temperature T, at which point they begin to lose their
magnetic properties and would drop from the magnetic field through the receiver once they reach
the desire temperature. This class of materials can also be magnetized by an external field, and
remain magnetized after the field is removed. For iron oxides, such as Fe2O4, these temperatures
can be as high as 1043 K [135].
Finally, electrically charged particles moving through a magnetic field tend to move in
circles orthogonal to the field, tracing out helical paths. An equivalent magnetic moment can be
assigned to the particle based on the time average of the particle’s circular motion. If the magnetic
field is non-uniform but has a small gradient, then the equivalent magnetic moment and the
resulting force from the non-uniform field are such that the charged particle, independent of the
sign of the charge, tends to be repelled by regions of strong magnetic field [135].

Figure 3.26. Particle temperatures (K) with downward particle acceleration of 9.81 m/s2 (left), 5
m/s (middle) and 2 m/s2 (right) [135].

69
In a study by Armijo et al. [135], the authors investigated high magnetic field strength materials
such as bismuth ferrite with melting temperatures >1100 ᵒC were assessed. The authors produced
a 3D DPM model for solid-particle flow employed where the ambient air is treated as a continuous
fluid (primary phase), and the dispersed solids (secondary phase) as an interpenetrating continua.
Particle-particle interactions were accounted for based on kinetic theory of granular flow. The
results of the investigators suggested that if induced magnetic fields are capable of reducing
particle acceleration from 9.81 m/s2 to 2 m/s2, the average particle velocity at the bottom of the
receiver is reduced from approximately 10-4.7 m/s [135]. This has the potential to increase the
average particle temperatures in this location from 449 K to 520 K (Figs. 3.26 and 3.27).

560

540

520
Temperature (K)

500 Gravity = 2 m/s^2

Gravity = 5 m/s^2
480 Gravity = 9.81 m/s^2

460

440
0 50 100 150 200 250 300
Particle Number
Figure 3.27. Particle temperature at bottom outlet for three Fluent runs with particle accelerations
of 2, 5, and 9.81 m/s2 [135].

3.5.1.4 Other Advanced Direct Receivers

Advanced direct receiver designs have been examined over a broad spectrum of other designs. The
use of a quartz window for cavity receivers has been investigated to reduce convective heat losses,
but potentially at the expense of reflected solar losses [140]. Quartz windows also limit the ability
to use pressurized gases and limit the size of the cavity opening [140]. Additionally, multi-cavity
receivers and microchannel designs are being pursued to increase the available heat-transfer area
to high-temperature, high-pressure fluids and media in compact designs, with the goal of reducing
heat loss and increasing thermal efficiency [141]. Additionally, extended heat-transfer surfaces
within tubes can increase heat transfer and structural reliability [142] and are being investigated
for high-temperature, high-pressure applications such as direct supercritical carbon dioxide sCO2
receivers [143]. These higher receiver temperatures are being investigated to enable higher-
efficiency power cycles such as sCO2 closed Brayton cycles. However, at higher temperatures,
radiative and convective losses from the receiver are increased. Heat transfer challenges also exist
across walls and surface interfaces during heat-transfer media transport to the storage system or
power block.

70
3.5.2 Indirect Absorption Receivers

3.5.2.1 Indirect Absorption Particle Receivers

For instances where heat is transferred through walls of a duct from a solar heat source to a gas,
heat transfer to the gas may be enhanced if the gas is seeded with small particles, where the
particles are at a higher temperature. This effect can be significant in cases where gases, which
have low heat carrying capacity, must be used in favor of high heat capacity liquids, due to high
temperature, reliability or other practical constraints. An optimization is then required for
transferring as much heat as possible to the gas to limit the necessary size of the heat transfer
surfaces.
Principally, for gas flowing through a tube without particles, heat is only transferred by
convection from the tube walls to the gas molecules. If the gas is radiatively participating, surface
emission from the interior of the tube walls is also effective in heating the gas. The presence of
particles therefore contributes to a third heat transfer mechanism where heat is transferred from
the particles to the gas by radiation, conduction and convection. In addition, heat is transferred
from the walls to the particles by radiation, especially if the radiative absorption coefficient of the
gas is low [138]. Heat is then passed on to the gas from the heated particles. Jones developed a
heat transfer model to study heated particles that are seeded in a heat exchanger running through
a furnace in order to enhance heat transfer to a gas flowing through the tubes. It was found that the
inclusion of particles heated to temperatures approaching the furnace temperature, and mass
loading ratios (seeded particle specific mass to carrying gas specific mass) up to an order of ten,
facilitated significant heat transfer enhancement from the walls to the gas [138]. Such enhancement
had the effect of reducing the required heat exchanger tube length and furnace size for the same
overall heat transfer.
Knowledge of temperature profiles within a packed bed and the heat removal rate is essential
for achieving particle flow homogeneity and optimal concentrating solar power (CSP) productivity
within channels. Predictive modeling for packed bed thermo-hydraulics is very complex due to the
interaction of different heat and mass transfer mechanisms on different scales and geometries
[144]. The study of heat transfer mechanisms between a packing material and a channel wall, of
a particle-gas fluid passing through a single tube, has traditionally been an area of research for
nuclear packed bed reactors and combustion research [145, 146]. For packed bed particle
configurations, inter-particle collisions are important thermal-physical phenomena that must be
adequately characterized for accurate heat transfer predictions, especially at high temperatures
[147]. Three approaches have been employed that address inter-particle collisions [148] for packed
beds. Again, as mentioned previously in section 3.5.1.2, these can include Eulerian-Eularian
models [149-153], Eularian-Lagrangian models [154-156], in addition to stochastic collision
models [157, 158]. Particle collision has been shown to affect the particle velocity field [159, 160]
and heat transfer in gas-particle flows. However, these models, which are considered as a standard
framework for investigating gas-particle heat transfer, have not to-date been validated for
temperatures as high as 720 ᵒC: the temperature SunShot goal for CSP [7]. Experimentally, for
packed bed heat exchangers, previous fundamental heat transfer research have only examined
operating temperatures up to 400 ᵒC for gas-particle flows containing alumina spheres, with
Reynolds numbers between 30-150 [161]. Hydrodynamically, although research has been
performed for Reynolds numbers as high as 10400, investigations to date have not examined
packed bed heat transfer beyond 12 MPa operating pressures [144]. Solid particles as an HTF

71
have been shown to reach operating temperatures as high as 750 ᵒC and flux densities as high as
250 kW/m2 [162]. However, an understanding of thermal-fluid transport phenomena of particle
beds for CSP applications is still relatively limited, especially for high temperature CSP particle
receivers. Accurate fundamental characterization of the particle-gas flow, as well as the particles
themselves, is necessary to improve heat transfer performance between the bulk fluid and heat
exchange walls, as well as to ensure reliability of the system for long-duration, nominal reliability
operation.
Interphase heat transfer between dispersed particles and a semi-continuous gas within a
channel has been studied by Jones [138] by formally modeling combined radiation, conduction,
and convection heat transfer between a particle and a semi-infinite medium. His results
demonstrated that with seeding particles, simple correlations for combined mode heat transfer can
be accurate. It was also found that the critical particle spacing at which interphase heat transfer is
interfered with by neighboring particles is smaller for radiation dominated cases than for
conduction dominated cases. Additionally, particle-particle and particle-wall collisions have also
been shown by investigators [164] to impact both the gas and bulk flow temperatures. Simulation
results by Mansoori et al. [164] indicate that the level of thermal turbulence intensity and heat
transfer are strongly affected by particle collisions. Inter-particle collisions were found to attenuate
the thermal intensity near the wall but somewhat amplify temperature fluctuations in the pipe core
region. The hydrodynamic-to-thermal time-scale ratio and the turbulent Prandtl number in the
region near the wall were also found to increase due to the inter-particle collisions. Results later
found by Mansoori et al. [163] showed that the effect of particle-particle heat transfer was more
significant for smaller particle sizes, lower flow Reynolds numbers, and for higher loading ratios.
It was also demonstrated that the particle volume concentration significantly affects the ratio of
inter-particle to gas-particle heat transfer.
For cases where interphase heat transfer is very rapid with small interphase temperature
differences, the gas/particle medium is often considered to be a single homogeneous phase. This
is the case for very small particles at the micron scale, especially when the initial temperatures of
the phases are the same [98]. Due to these fine particle sizes, studies of such cases tend to
concentrate on turbulent particle transport and the effect on the total medium radiative properties
of the distribution of particle number density and size [138]. For turbulent gas-solid particle flow
regimes, experimental evidence suggests that the Nussselt number is increased by adding hot
particles to a fluidized bed with large particles and low gas velocities, while it can be decreased
when the particles are small or the gas velocity is high. Research by Mansoori et al. [165] validated
this finding with particles that ranged between 600-1200 μm and velocities between 11-21 m/s.
Their research found for flow within the turbulent regime that a Lagrangian-Eulerian modelling
framework was able to show agreement with their experimental results for the particle sizes and
flow velocities considered.
Advances have been made in characterizing different particulate flow regimes and their
respective transitions, which impact thermal convection between the bulk gas-particle fluid and
the heat exchange walls. For these systems, the analogy with gas-liquid two-phase flow can be
applied for investigating hydrodynamic regimes in gas-solid systems [166]. If particle flow is
dispersed, such as the case of particle separation being greater than 10 diameters, turbulent
conditions are then considered along with particle mixing theory [166]. Although issues still
remain pertaining to gas-particle instability theory based purely on hydrodynamic considerations,
most fluidization researchers believe that interparticle forces play an important role for powders
less than 100 μm [167]. However, work by Rosenweig [168], demonstrated that limiting bubble-

72
free hydrodynamics was important for increasing heat transfer. In their vertical particle-gas
hydrodynamic studies, Flammant et al. [162] demonstrated that dense particle flows (with SiC
particles having a solid fraction in the range of 30-40%), with optimally low-levels of air voids or
“bubbles,” help to facilitate homogeneous temperature distributions between the particles and the
continuum gas by enhancing agitation and thermal mixing, which is favorable for heat transfer
[167].
For some thermochemical processes heat transfer in two-phase systems involving gas and
solid particles are encountered, fluidized beds operate in either the bubbling regime or the fast-
circulating regime. In the first case, particles are retained in the fluidized bed while gas passes
upward past the particles, partially as rising bubbles. In the second case, gas velocities exceed
terminal velocity for the individual particles and the two phases flow through the fluidized bed in
concurrent upward flow [169]. As the velocity increases various hydrodynamic regimes of gas-
particle fluid flow occur, which are presented in Fig. 3.28.

Figure 3.28. Hydrodynamic regimes of fluidization [170].

From Fig. 3.28 increased flow facilitates dispersion (“bubbling”) between the particles and
increased turbulence, where Wen and Yu 171] developed Eqn. 3.50 to estimate the minimum fluid
velocity for spherical particles. Their results suggested that increasing gas velocity beyond
minimum fluidization causes the excess gas to collect into discrete bubbles that grow and rise
through the fluidized matrix of solid particles.
𝑈𝑚𝑓 𝐷𝑝 𝜌𝑔
𝑅𝑒𝑚𝑓 = = [(33.7)2 + 0.0408𝐴𝑟]1⁄2 (3.50)
𝜇𝑔

𝐷𝑝3 𝜌𝑔 (𝜌𝑠 −𝜌𝑔 )𝑔

𝐴𝑟 = (3.51)
𝜇𝑔

Typically, tubes carrying HTFs are often immersed in fluidized beds to extract or add thermal
energy [167]. Data for the circumferentially averaged heat transfer coefficient for horizontal tubes
are shown in Figure 3.29 for various types of solid particles.

73
Figure 3.29. Average heat transfer for horizontal tubes immersed in bubbling fluidized beds for
various solid particles [172].

Additionally, Kunii and Levenspiel [127] have shown that increasing gas pressure and density
significantly increases the magnitude of the heat transfer coefficient as well as promoting the
occurrence of minimum fluidization at a lower value of superficial gas velocity. The effect of
bundle spacing is insignificant at 1-atm pressure but becomes increasingly more important as gas
pressure and density increase [167].
For fast fluidization, when the superficial gas velocity exceeds the terminal velocity of the
solid particles, the particles are generally suspended in concurrent upward flow with the gas. This
upward flow occurs in rise reactors where desired physical or chemical reactions occur. In most
applications, the two-phase flow exits the top of the riser into a cyclone where the gas phase is
separated and exhausted while the solid particles are captured and returned for reinjection at the
bottom of the riser [167]. Volumetric concentration of solid particles in fast fluidized beds (FFBs)
tends to be fairly diluting [127] however. Heat exchange with the particle/gas suspension is usually
accomplished through the vertical wall surfaces or through vertical tubes immersed in the duct.
The heat transfer coefficient on the vertical surfaces of FFBs has been found to increase with
increasing solid concentration, aside from other second-order parametric effects. It is generally
accepted that the effective heat transfer coefficient for surfaces in FFBs has contributions for gas-
phase convection, particle-induced convection, and radiation as prescribed by Eqn. 3.52.

ℎ = ℎ𝑔 + ℎ𝑝 + ℎ𝑟 (3.52)

In contrast to the situation in dense-bubbling fluidized beds, the relatively dilute concentration of
solid particles in FFBs often results in significant contributions from all three heat transfer
mechanisms [127]. The radiation coefficient can be obtained by a gray body model suggested by
Grace [167]. The contribution of the gas phase convection (hg) is commonly estimated based on
correlations for gas flow alone at the same superficial gas velocity. Although the presence of
particles may alter the turbulence characteristic of this gas flow, any errors caused by this
procedure are usually small since hg is generally smaller than the particle-phase convective
coefficient hp. For most FFBs, the particle convective contribution to heat transfer is most

74
important and the prediction of hp is the major concern in thermal design [167]. Unfortunately,
mechanistically based models are still lacking and most design methods rely on empirical
correlations which often combine the contributions of gas and particle phases into a single
convective heat transfer coefficient (hc). One such correlation proposed by [173] where Vt is the
terminal velocity of the particles:
0.3 0.21
ℎ𝑐 𝐷𝑝 𝐶𝑝𝑝 𝜌𝑠𝑢𝑠 𝑉𝑡
𝑁𝑢𝐷,𝑝 = = (𝐶 ) ( 𝜌 ) (𝑔𝐷 ) 𝑃𝑟𝑔 (3.53)
𝑘𝑔 𝑝𝑔 𝑝 𝑝

For high thermal input conditions, much higher values of gas flow than predicted by purely
hydrodynamic approaches are required when the temperature is raised to a point where
agglomeration occurs [174]. However, as agglomeration intensifies, convection heat transfer
coefficients have been shown to reduce [175] due to decreased particulate mixing. However,
external fluidization techniques such as imposed acoustic fields [176] and magnetic fields [168,
178] can cause disaggregation of particle clusters into smaller units composed of a few relatively
large agglomerates and many smaller ones, allowing in some cases large particle sizes of up to 45
μm to be fluidized as if they were smaller materials [176, 177]. Research by Chirone et al. [176,
177] indicated that acoustical field effects extend to gas velocities well above minimum
fluidization and result in much more reproducible behavior of cohesive powders.

3.5.2.2 Tubular Receivers

Tubular receiver designs have previously been employed by parabolic dish receivers as well as
liquid–metal heat pipes for improving heat exchange from a solar heat input to a gas [179]. The
internal heat transfer coefficient in a liquid-metal heat pipe is on the order of 30,000W/m2-K
compared to 300W/m2-K for heat transfer to gases [179]. Therefore, higher solar fluxes can be
tolerated with heat pipes yielding more compact receivers, lower metal temperatures, and lower
pressure drops where disadvantages include potentially higher receiver costs [64]. Design
specifications included an air-outlet temperature of 815 ᵒC with an air-inlet temperature of 565 ᵒC,
air mass flow rate of 0.24kg/s, pressure drop of 2%, and thermal efficiencies up to 85% [180].

Figure 3.30. SOLGATE tubular receiver for 600 ᵒC air outlet temperature [180].

More advanced designs for this type of system have been conducted by DLR for a solar-hybrid
microturbine system operating in a central receiver for applications on the order of 100kW–1 MW
[181-183]. In one study, investigators as part of the SOLGATE project, developed a solar-hybrid
cogeneration system with a micro-turbine, which was able to operate under varying contributions

75
of solar power input and fuel. A novel profiled multilayer tubular receiver, shown in Fig. 3.30,
demonstrated operation at a power level of 230 kWe and temperatures of approximately 600 ᵒC
[180].
During development of a low temperature tube module, analyses by Heller [180] were
conducted to evaluate the upper limit of a tubular receiver for solar hybrid gas turbine systems.
Typically, there is a strongly inhomogeneous heat load on the absorber tubes, with solar fluxes up
to 500 kW/m2 on the front side and low fluxes on the backside of the tubes [180]. Two limiting
factors were identified: 1. low heat conductivity of high temperature metal alloys and 2. poor ratio
between heat transfer coefficients and pressure drop. Finite element calculations of the temperature
and stress distributions at the cross-section of the absorber tubes demonstrated high thermal
gradients in the tube walls (up to 220 K), caused by low thermal conductivity of the tube material
and relatively low heat transfer coefficients between the tube wall and air at the inner side of the
tube [64]. This led to high inhomogeneous stresses, reduced life time, as well as a relatively low
air outlet temperature compared to the maximum allowable surface temperature [180]. To reduce
thermal gradients between the irradiated front side and the non-irradiated back side of the absorber
tubes the concept of a “profiled multilayer tube” (PML) technology was developed, which consists
of using three metallic tube layers: a high temperature alloy at the outer side, copper as
intermediate layer and another high temperature alloy layer at the inner side of the tube, Fig. 3.31.

a. b.

Figure 3.31. a. Profiled multilayered tube b. PML tube sample of normal steel and copper [180].

The function of the copper within the PML is to leverage its heat conductivity to distribute the heat
by conduction throughout the circumference. The outer layer however provides the structural
strength while the inner layer protects the copper from corrosion at high temperatures. In addition,
a spiral profile was implemented on the inner surface to improve the heat transfer coefficient.
Preliminary calculations for this solar receiver concept revealed that the temperature gradient in
the tube wall could be significantly reduced, in addition to the maximum temperature of the tube
[180]. Figs 3.32a and 3.32b presents the simulated temperature distribution in standard and multi-
layer absorber tubes respectively.

76
 a. b.
Figure 3.32. Temperature distribution in an irradiated a. standard tube and b. multi-layer tube
[180].

The empirically validated results in Fig. 3.33 demonstrated that the temperature difference
between the irradiated and non-irradiated sides of the tube section can be reduced from 106 °C to
31 °C when utilizing PML tubes. Therefore, the investigators concluded that the application of
PML tubes instead of common mono tubes leads to a noticeable decrease of the maximal overall
temperature and also to a significant reduction of the temperature gradient relating to the tube
circumference and, hence, to a reduction of the internal stresses [180].

Figure 3.33. Local temperature distribution comparisons of a mono-tube and a PML tube [180].

Tubular receivers however have many challenges associated with large convective and radiative
heat losses from receivers operating at higher temperatures, as well as difficulty in transferring
heat effectively from irradiated tubes to the gas. They are also subject to rapid transient
thermomechanical loads that can adversely affect the fatigue life of the receiver [64]. Uhlig
conducted tests and developed a Chaboche-type plasticity model to predict the fatigue life of
nickel-based alloy tubes subject to transient thermal stresses, where the model was used to reduce
the stresses in the receiver [184]. Kolb compiled low-cycle fatigue data for Incoloy 800HT, Inconel
625-LCF, and Haynes 230 alloys and performed analyses to determine allowable flux limits on
these materials [185]. Tubular receivers are a likely possibility for employment in supercritical
carbon dioxide (sCO2) systems as small diameter tubes can enable high pressures required for the
supercritical phase. At the turbine inlet, pressures of approximately 15–25 MPa may be expected
for sCO2 Brayton cycles [64].

77
3.5.2.3 Volumetric Receivers

Air-cooled tubular receivers can experience challenges due to lower heat transfer coefficients, as
discovered in the German–Spanish GAST project where two tubular receivers, one metal and one
ceramic, were tested at the PSA in Spain [186]. To improve the contact surface, volumetric
absorbers with very porous wire, foam or appropriately shaped materials within a volume, allow
radiation to penetrate deeply into the structure where thin substructures (wires, walls or struts)
ensure good convective heat transfer [186]. A well-designed volumetric absorber produces the so-
called “volumetric effect” as the irradiated side of the absorber is at a lower temperature than the
fluid leaving the absorber. Under specific operating conditions, volumetric absorbers tend to have
an unstable mass flow distribution [187]. Receiver arrangements with mass flow adaptation
elements (e.g., perforated plates) located behind the absorber can reduce this tendency, as well as
appropriate selection of operating conditions and the absorber material [187].
During operation, the HTM is heated to temperatures between 800-1000 ᵒC for metals, up
to 1200 ᵒC for ceramics, and up to 1500 ᵒC for SiC [188]. The HTM can then be used to heat
another working fluid for a power block [189], charge a storage medium [190], or pass directly
into a gas turbine [64]. Two prominent applications of volumetric air receivers provided by Ho
and Iverson [64] are a. open-loop atmospheric receiver system and b. closed-loop pressurized
(windowed) receiver system for a Brayton Cycle. A comprehensive review of volumetric receivers
can be found by Avila-Marin [188].
Losses due to radiation can be mitigated by a two-slab selective volumetric receiver in
which the irradiated front slab was composed of solar transparent glass beads or a silica
honeycomb, and the second slab was composed of silicon carbide particles [191]. These
semitransparent multi-layer systems allow solar radiation to penetrate through the first slab and
into the second slab. Infrared emission from the second slab can be absorbed by both slabs and as
a result, the location of maximum temperature can be found within the interior of the volumetric
absorber, therefore decreasing radiative heat loss. Experiments by Avila-Marin [189] found for a
windowed receiver, thermal efficiencies of approximately 90% could be attained with gas outlet
temperatures of approximately 700 ᵒC could be attained. Pitz-Paal et al. [192] also suggested a
similar concept utilizing square glass channels that cover a ceramic foil receiver where their results
suggested that the efficiency could be improved by up to 10% with gas outlet temperatures up to
1000 ᵒC.
The potential for unstable flow and non-uniform heating in the volumetric receiver, leading
to overheating and local failures in the receiver material [64] is another challenge for volumetric
air receivers [193]. These instabilities are caused by changes in temperature-dependent air
properties (viscosity and density), but may be mitigated by using low-porosity absorber materials
[64]. Karni et al. suggested the use of a volumetric solar receiver employing ceramic pins which
helped achieve gas temperatures near 1000 ᵒC [194].

78
4. HEAT TRANSFER FLUIDS FOR CSP SYSTEMS
4.1 Overview

To achieve high conversion efficiencies for CSP systems, high temperatures at the receiver and
heat engine, along with low temperatures at its energy sink, are inherently necessary. These
temperatures are limited by the HTFs (or heat transfer media (HTM)) in the receiver and the
working fluid in the heat engine. For some systems the HTF and working fluid can be the same,
as in the case of water/steam [195]. Additionally, the selection of an HTF must entail consideration
of its ability to retain and store energy, as this is an important feature for CSP. HTFs can be
classified by their states of matter during their respective operating conditions. In addition to the
three standard states (solids, liquids and gases), HTFs can also experience higher heat transfer
coefficient levels as they undergo phase change processes and achieve supercritical
thermodynamic states.
Previous work rated the potential of various HTFs for CSP applications by their thermal
and transport properties, where Becker [196] developed the following comparisons according to
their density Fig. 4.1a, specific heat Fig. 4.1b, dynamic viscosity Fig. 4.1c and thermal
conductivity Fig. 4.1d.

a. b.

c. d.

Figure 4.1. Selected CSP HTF comparisons according to a. density, b. volumetric heat capacity,
c. dynamic viscosity and d. thermal conductivity [196].

79
When combining respective properties from Fig. 4.1, the Prandtl number PR can be determined as
the ratio of the momentum diffusivity to the thermal diffusivity, where from Fig. 4.2, fluids such
as water, Hitex and air have Prandtl numbers greater than one, while alkali fluids, which were
found to achieve the highest operating temperatures, had values less than one [196].

Figure 4.2. Computed Pr values for selected CSP fluids. [196].

Heller [197] notes that for ideal CSP systems, important thermophysical properties of HTFs to
consider are:

 Low solidification temperature.

 High evaporation temperature and thermal stability limit.
 High material or bulk fluid thermal conductivity.
 Low viscosity which can contribute to low pumping power requirements.
 High densities and heat capacities which can facilitate ideal storage requirements.
 The ability for the fluid to be used as a power block working fluid to reduce exergy losses.
 Chemical compatibility with contact metals with low corrosion/oxidation activity values.
 Low cost and high availability
 Low toxicity, flammability, explosivity and other environmental hazards.

High performance HTFs have the requirement of gaining heat from the receiver but also efficiently
rejecting heat to a thermodynamic cycle. This heat transfer capability is based on convective heat
transfer characteristics, as outlined in section 2.3, which includes a high thermal conductivity that
enables efficient transfer of heat from the absorber to the power block, a high density and specific
heat capacity that enables high heat flux levels at reasonable mass flow rates, and low HTF
viscosities that minimize pumping power [139]. To capture these requirements, HTF figures of
merit (FOM) have been developed where Mouromtseff [295] proposed Eqn. 4.1 FOM to compare
the performance of HTFs based on the Dittus-Boelter correlation for internal turbulent flow.

𝜌0.8 𝐶𝑝0.33 𝑘 0.67

𝐹𝑂𝑀𝑀𝑜 = 𝜇 0.47
(4.1)

80
Other FOMs suggested by Weinstein et al. [139] include one by Bonilla et al. [296] for axial heat
flow, Eqn. 4.2 which excludes flow of heat from a wall, as well as an FOM by Lenert et al. [297]
Eqn. 4.3, which is based on the minimization of pumping power given a certain temperature rise
in the collector tube.
𝜌2 𝐶𝑝2.3
𝐹𝑂𝑀𝐵𝑜𝑛𝑖𝑙𝑙𝑎 = (4.2)
𝜇 0.2

𝜌2 𝐶𝑝2.3
𝐹𝑂𝑀𝐿𝑒𝑛𝑒𝑟𝑡 = (4.3)
𝜇 0.2

Weinstein et al. provided a comparison for these FOMs for several CSP HTFs where they found
FOMs by Mouromtseff and Lenert et al. had similar trends, as illustrated in Fig. 4.3. The values
where found highest for liquid metals due to their high thermal conductivities. Additionally,
saturated water was found to be very high for all FOMS due to its high heat capacity while air and
pressurized water vapor had the lowest FOMs due to their low viscosities [139].

a. b. c.

Figure 4.3. Several selected CSP HTF materials comparisons by Weinstein [139] for FOMs by a.
Mouromtseff [295], b. Bonilla et al. [296] and c. Lenert et al. [297].

Finally, with regard to the thermodynamics of HTFs, the fundamental equation of state can be
assessed to evaluate their ability to operate under high pressure conditions (e.g. for sCO2 and sH2O
systems). Traditional approaches for formulating equations of state are: 1. methods of mechanical
statistics, which take a rigorous approach for assessment of the interaction between molecules, and
2. empirical or semiempirical equations to describe their behavior based on experimentation. The
semiempirical equations of state generally express pressure as the sum of two terms, a repulsive
term PR and an attractive term PA. According to this model, no single molecule can move freely,
but interacts with nearby molecules via forces of cohesion and repulsion. For most analyses one
can utilize the van der Waals equation of state Eqn. 4.4, to describe the properties of a real gases
qualitatively [198].

81
 𝑅𝑇 Θ(𝑉−𝜂)
𝑃 = 𝑃𝑅 + 𝑃𝐴 = 𝑉−𝑏 − (𝑉−𝑏)(𝑉 2 +𝛿𝑉+𝜖) (4.4)

The acentric factor, ω was introduced by Pitzer et al. [199] as a measure of the difference in the
structure of a molecule of any substance compared to that of a gas with a spherical molecule
defined as:
𝜔 = −𝑙𝑜𝑔[𝑃𝑣𝑝,𝑟 ] 𝑇 =0.7 = −1 (4.5)
𝑥

where [𝑃𝑣𝑝,𝑟 ] 𝑇 = [𝑃𝑣𝑝 ] 𝑇 ⁄𝑃𝑐𝑟 which is the reduced vapor pressure calculated at the
𝑥 =0.7 𝑥 =0.7
reduced temperature Tr = 0.7. By introducing reduced variables 𝑇𝑟 = 𝑇⁄𝑇𝑐𝑟 , 𝑃𝑟 = 𝑃⁄𝑃𝑐𝑟 and 𝑉𝑟 =
𝑉 ⁄𝑉𝑐𝑟 , into the Van der Waals equation, on can arrive at the reduced pressure Eqn. 4.6, where the
compressibility factor at the critical point is defined by Eqn. 4.7.

𝑇𝑟 𝑎∗ 𝛼(𝜔,𝑇 )
𝑃𝑟 = 𝑍 − 𝑍 2 𝑉 2𝑐𝑟 𝑟
(4.6)
𝑐𝑟 𝑉𝑟 −𝑏 ∗ 𝑐𝑟 𝑟+2𝑏 ∗ 𝑍 (𝑏 ∗ )2
𝑐𝑟

𝑃𝑐𝑟 𝑉𝑐𝑟
𝑍𝑐𝑟 = (4.7)
𝑅𝑇𝑐𝑟

with 𝑏 ∗ = 0.077796, and 𝑎𝑐𝑟

∗
= 0.457235. Eqn. 4.3 corresponding to the critical point where Tr =
Pr = Vr =1, Zr has previously been found to be a constant 0.307 for any fluid [198]. In a study by
Intermizzi [198] the investigator found for all fluids tested, the same volumetric behavior except
for those with a varying acentric factor ω. The results can be seen below in Fig. 4.4 for various
working fluids tested.

Figure 4.4. Working fluids compressibility chart where the values of the compressibility factor Z
are calculated from the Van der Waals equation of state [198].

Overall, the sole purpose of an HTF is to transmit heat from one location to another. There are
very diverse types of HTFs for CSP technologies that are typically employed based on their
thermodynamic property limits and costs. A CSP design consideration that impacts these property
limits is whether the HTF is directly or indirectly heated. As illustrated Figs. 4.1-4.4, most
thermodynamic properties are strongly dependent on temperature and pressure and can vary
significantly if its fluid motion is bounded or unbounded. Additionally, since many CSP
technologies may also have inherent optical concentration constraints, relative operating

82
temperatures can also be assessed, in addition to costs, to determine the best HTF for a particular
application. In this chapter, various HTF fluids, spanning various phases of matter will be assessed
for CSP applications, temperatures and pressure conditions.

4.2 Synthetic Oils

Parabolic trough collectors capture and focus solar radiation onto a linear receiver tube that is
placed in the focal line of the parabolic trough collectors. The HTF in these systems is typically a
synthetic oil [200], which circulates through the receiver tube and transports the thermal energy to
a heat exchanger system in a power block, where high-pressure superheated steam is generated
[452]. A receiver tube loop runs through the solar field and is composed of a modular series of
heat collection elements (HCE), which include steel tubes that contain synthetic oil with an outer
glass tube that maintains a vacuum around the hot steel tube. This vacuum annulus serves as a
thermal insulator while allowing for maximum absorption of the solar thermal radiation by each
HCE [452].
Common synthetic oil HTFs have previously included mineral oil, silicone, aromatics and
polyalkylene glycols due to their employment at relatively low pressures [433]. However, in most
cases synthetic oils are composed of a eutectic mixture of 26.5% Biphenil and 73.5% Diphenil
Oxide [432], with commercial names such as Therminal VP-1® or Dowtherm® synthetic oils [197].
These fluids have stable molecules that remain liquid under low pressures from ambient
temperatures up to about 400 °C [201]. The most commonly used grades of these HTFs in CSP
plants typically have a purity as high as 99.9%, with chlorine levels below the detection limit of
0.2 parts per million (ppm). Disadvantages for employment of these oils are degradation reliability
issues and high costs. Overall, since their upper limit is relatively low, they are generally unable
to be used in in high efficiency power cycles [197]. Additionally, previous work has also identified
degradation mechanisms that can limit their utility over time where specific synthetic oil grades
with high chloride and sulfur levels may not be suitable for use in CSP plants due to the risk of
leaks and expensive equipment replacements from stress corrosion cracking. Lang and Lee [432]
indicated that thermal degradation is impacted by many factors which include:

 Fluid temperature
 Initial impurity concentration (organic and inorganic)
 Concentration of degradation products in HTF
 Low boiler / high boiler ratio in the HTF
 Degradation product composition
 Oxygen accessed to the fluid (e.g. lack of cover gas quality)
 Metal surface to HTF mass ratio due to catalytic effects from construction materials
 Contamination of HTF (e.g. pipe conservation oil, residues from construction, oil from pump
seal system, etc.)

In their work, the investigators developed a model (Eqn. 4.8) for thermal degradation, k of the
eutectic diphenyl oxide/biphenyl fluids in a parabolic trough CSP plant was built based on
laboratory testing results and average fluid analysis results from operating CSP plants.
2 +𝑏𝑇+𝑐
𝑘 = 𝑒 𝑎𝑇 (4.8)

83
From Eqn. 4.8, T is the temperature of the fluid where the three coefficients are determined
empirically with respect to the fluid concentration. Fluid degradation was compared between a
high quality product at 99.9% purity and commercially available products with reduced purities of
around 99.5%. A range of 1.7-2.4 times more degradation was determined for 99.5% purity fluids
compared to the 99.9% quality for the operating conditions in a CSP plant. These results indicate
that more frequent degradation separation mitigation operations would be needed for lower 99.5%
purity fluids, which ultimately would require higher costs where, additional new fluid refills over
a 25 years of operation would be approximately $2/kg [437].
Another issue using synthetic oils as an HTF is the formation of hydrogen gas over time as
a result of thermal degradation, which proceeds at a very low rate at 400°C, but accelerates
considerably between 400°C and 425°C [452]. H2 forms within steel receiver tubes and permeates
into a vacuum annulus over a very long time period, where it can facilitate heat loss. Research by
Moens and Blake [452] provided mechanistic evidence for the degradation process which involves
a radical reaction where hydrogen gas is formed as a by-product, as well as aromatic oligomers
with higher molecular weights. Additionally, the authors also suggest that organic impurities in
the fluid can catalyze the thermal breakdown of the HTF through the thermal formation of highly
reactive hydrogen atoms (radicals). Here, the authors recommend the HTF should be of the purest
grade and kept free of organic impurities to mitigate hydrogen formation [452]. Another
contribution to hydrogen formation is the presence of oxide layers at the surface of the steel
receiver walls that can have sufficient catalytic activity to convert DPO into dibenzofuran. Further
research is still required however to understand the fundamental chemistry behind this catalytic
reaction which has never been thoroughly investigated [452].

4.3 Molten Salts

The biggest advantage of molten salt as a HTF is the possibility of direct storage at relatively low
costs. Molten salt power tower systems are generally limited to receiver outlet temperatures of
approximately 565-600 ᵒC and employ heat exchangers to produce superheated steam at 540 ᵒC
and 130 bar, as well as reheated steam at 538 ᵒC and 28 bar [202]. The resulting gross thermal
efficiencies are typically 42% with wet cooling [203].
Commonly used nitrate salts (60% NaNO3 and 40% KNO3 by mass) are stable to higher
temperatures of up to 600 ᵒC their lower maximum temperature can be limiting as further increased
temperatures increases the corrosion rate with stainless steel [204]. Generally, heat transfer
characteristics of molten salts are mediocre where a reasonably high density and low specific heat
capacity enable a low volume flow but the low thermal conductivity can lead to elevated thermal
losses [197]. To reduce these losses and increase receiver efficiency, heat transfer between a pipe
and HTF can be improved by increasing fluid velocity and turbulence [205], while optical
efficiency can be improved with pyramid-like spikes of which the HTF circulates [206].
The high solidification temperatures of liquid salts can present challenges, especially with
linear receivers where the HTF could freeze during evening hours or during times of low irradiance
[197]. In central receiver systems, salt will normally be drained into a tank during receiver filling
with a cover gas, though the freezing of the salt, which could block pipes and valves, could cause
severe damage to a system [207]. Just as with molten metals, solutions to freezing, aside from
draining, could include trace heating or circulation of stored salts. However, these solutions could

84
result in higher heat losses, electrical power consumption and investment costs. Research has been
conducted to address this issue for the development of salts with lower melting points [197], as
well as higher maximum operating temperatures to facilitate higher efficiency power cycles [207].
Kolb et al [185] estimated the economic benefits of raising the salt HTF’s receiver exit temperature
to approximately the same temperatures as investigated by Kelly [208], and predicted LCOE
reductions of up to 8%. Advanced salt concentrations have in turn been more recently studied in
an effort to raise the power cycle operating temperatures. Raade et al. [209] found a quinary
composition of LiCl, NaCl, KCl, CsCl and SrCl to have a melting point of 253 °C at ambient
pressure, and a thermal stability of up to approximately 750 ᵒC. This novel salt consisted of more
than 70% by weight of CsCl and LiCl, which however are both expensive materials.
Earlier studies have previously been performed to understand molten salt thermal-fluid
behavior in advanced CSP molten salt reactor applications [211, 212], where the majority of the
literature on molten salt heat transfer studies have been limited to relatively low Re numbers less
than 45000 [210]. Hoffman [213] investigated turbulent heat transfer for molten NaOH for Re
values of 6000-12000 where the authors determined heat transfer characteristics followed the
correlation:
𝑁𝑢 = 0.021𝑅𝑒 0.8 𝑃𝑟 0.4 (4.9)

A wider range of Re was studied by Grele and Gedeon [214] between 5300-30000, where the
authors measured values of Nu were 20% higher than those of the widely-used Dittus-Boelter
equation, Eqn. 2.33. Hoffman and Cohen [215] studied molten NaNO2-NaNO3-KNO3 for heat
transfer for Re between 4850-24710 and between heat flux up to 614 kW/m2, where the authors
also found good agreement between their data and the Dittus-Boelter equation. In another study,
Yu-Ting et al. [216] researched molten LiNO3 salt flowing through a test section that rejected heat
to a cooling oil flowing a jacket around a test section with a Re range 4100-9850 where the authors
showed good agreement between their experimental data and the Gnielinski equation [217] and
the Hausen equation [216]. Additionally, a study by Das et al. investigated molten nitrate salt [210]
(60% NaNO3 and 40% KNO3) up to Re values of 2x105 and heat flux levels up to 1,000 kW/m2.
The authors found good agreement between their developed Nu correlation and their data, where
two major effects were observed: 1. the Nu values gradually plateaued for the highest Re tested
and 2. at higher Re values, positive interaction of heat flux on Nu was observed [210].
Salts containing lithium have also been studied for thermal energy storage (TES) systems
applications because of their optimal thermophysical properties. Lithium nitrate has been
considered as a great additive to improve the thermal performance for sensible heat storage of
molten salts, due to its ability to increase the salt mixture’s working temperature range [435]. In
their research, Fernández et al. [436] characterized thermophysical properties of lithium nitrate
containing salts with corrosion evaluations of carbon and low chromium steels at 390°C for 1000
hours. Thermophysical properties, such as melting point, heat capacity and thermal stability of
saline nitrates were measured using a simultaneous thermogravimetric analyzer (TGA) and
differential scanning calorimeter (DSC), which included NaNO3 and KNO3 (SQM-SSR grade),
Ca(NO3)2*4H2O and LiNO3 (Panreac 98%) [436]. Their results, as shown in Table 4.1, showed
that using lithium and calcium nitrate, in small quantities (10 wt.%), enabled the salt mixture
energy density to increase by 19%. Their results also found a reduction in thermal stability due to
the incorporation of corrosion products into the salt. A steep reduction in weight % was observed
around 350 °C in the salts after the corrosion tests.

85
 Table 4.1. Thermophysical properties of lithium nitrate containing molten salts [436].

To address corrosion issues, especially with chloride salts which are extremely corrosive to
stainless steels, nickel alloys can be employed, however their costs can be prohibitive. However,
chemically resistant coatings can also be applied to ensure system internal surfaces have low
wettability and can withstand long-term exposure to these materials at temperatures greater than
600 °C [440]. When power turbines and containment alloys reach temperatures greater than 700
°C, the salts form a thin film of molten fluid that attacks the substrate metal alloy (hot corrosion).
The applied coatings are designed to limit degradation of the underlying substrate [437, 438] where
they must be very dense with minimal defects during application to ensure good resistance [439].
Protective coatings are formulated to corrode at an established rate (µm/yr. normal to the surface)
when in contact with the HTF. This corrosion rate determines the coating thickness that is required
to protect the substrate for a 30-year lifetime of a respective plant. To protect storage tank walls
and piping from corrosion and eventual failure, coatings are being developed that have a target
corrosion rate of 30 µm/yr. with a minimal thickness of 900 µm [440].
Chloride salts have not been observed to form stable passivated oxide layers and are
susceptible to high corrosion rates in the presence of moisture and oxygen. Purification techniques
however can be employed to ensure formation of protective layers do not readily occur where
ullage gas may also be employed to reduce corrosion, though a study by Kruzenga [441] showed
that oxygen content must be lower than 10%. Impurities in the form of oxygen and moisture are
much less of an issue in carbonate salts. For carbonate salts ullage gases are typically high in CO2
content, where formation of protective oxide layers has been found to inhibit corrosion from the
base alloy [441]. Additionally, Ren et al. [454] and Olivares et al. [455] proposed replacing molten
nitrate salts with mixed carbonate salts since their corrosion-reducing ability could further increase
operation temperatures to between 700 °C and 850 °C. Research by de Miguel et al. [453]
investigated corrosion behavior of an austenitic steel HR3C in a eutectic ternary carbonate molten
salt mixture (Na2CO3/K2CO3/Li2CO3) through an isothermal immersion test at 700 °C for 2000
hrs. A microstructural and compositional study was performed using SEM – EDX and XRD
analysis where they observed corrosion products arranged in a multilayer structure, with LiFeO 2,
LiCrO2, NiO and FeCr2O4 being the main compounds in different layers (from molten salt to the

86
unaffected substrate interface). The weight loss observed through gravimetric analysis was
attributed to soluble chromates initially formed (K2CrO4). Despite this weight loss the utility of
carbonate steels in CSP applications still appears promising, however further research over longer
time periods at high temperatures (~700 °C) is still required to ensure compatibility of carbonate
salts that the HR3C steels [453], as well as other CSP materials.

4.4 Metals

Traditionally, latent metal HTFs typically don’t have the problematic upper, and lower operating
temperature limitations as experienced molten salts. Latent metal HTFs can operate at lower and
safer pressures, while still achieving temperatures required for next-generation Rankine, Brayton
and other power cycles. These materials can possess outstanding heat transfer with low viscosity
and high thermal conductivities which can facilitate small thermal gradients of the flow inside the
receiver pipe wall. Additionally, the pipe thickness can be kept small as well due to the low
pressure, which leads to maximum pipe temperatures close to the fluid’s exit temperatures [222],
resulting in higher receiver efficiencies and reduced strain inside the pipe caused by thermal
expansion [223]. Eventually, these properties may allow higher maximum solar fluxes on the
receiver and a higher thermal efficiency of it. Kelly [208] determined the maximum allowable
incident flux on a molten salt receiver with an operating pressure of 20 bar and a 696 ᵒC outer max
pipe temperature, to be as high as 2.5 MWth. It is believed however that liquid metal receivers can
tolerate even higher influx levels [197].
Thermodynamically, latent metal receiver systems have the potential for having an
enhanced exergetic fit to higher temperature CSP systems as shown by Fig. 4.5 [218]. Here
Andraka et al. demonstrated that latent HTF CSP engines have significantly reduced exergy losses
compared to sensible systems, especially when combined with a latent storage system [226].
Liquid metals have garnered growing attention as a potential replacement for more conventional
HTFs in CSP systems due to their high thermal conductivity, which have a strong potential for
increasing receiver efficiency and operational temperatures that could facilitate more advanced
power cycles. They also have very high potential as HTFs due to their wide range of practical
operating temperatures and their superior heat transfer characteristics.

Figure 4.5. Comparison of sensible and latent heat transport into a Stirling engine [219].

87
Latent metal systems also allow for high maximum fluid temperatures at low pressures, high
receiver efficiencies and low pressure drops. For binary cycles the utility of alkali metals as a
working fluid for topping cycles (as shown in Fig. 4.6 for saturation pressure curves [220]) are
generally considered as a means of increasing the operating temperature beyond that of standard
steam cycles. As shown in Table 4.2. Liquid-metal HTFs can facilitate high absorption from very
high solar flux levels and efficiencies where the utility of heat pipes is also being evaluated as a
way to increase heat transfer from receiver tube or panel surfaces.

Table 4.2. Peak Flux comparison of common CSP HTFs with liquid Sodium [221].

In conjunction with a combined cycle the use of alkali metal HTFs in latent form can also improve
performance of Rankine power cycles. As illustrated by Fig. 4.6, an analysis performed by
Angelino and Invernizzi [220] at a maximum temperature equal to that of steam, potassium was
shown to dramatically reduce major thermodynamic limitations of the steam cycle by increasing
the overall average temperature of the heat input from the primary source.

Figure 4.6. Conceptual topping cycle comparison between potassium and steam at the same
maximum temperature [220].

Alkali metals have been proposed for high-temperature topping cycles in Rankine-Rankine
combined cycle configurations [220], where these systems have the potential to provide good
temperature-matching with bottoming steam cycles due to higher phase change heat transfer
coefficients. Thermal efficiencies of an alkali Rankine topping cycle and steam Rankine bottoming
cycle have however been shown to reach 56% for 1000 ᵒC maximum HTF temperatures, with the
potential of reaching as high as 60% [203]. However, turbomachinery for the metal vapor cycles
is estimated to be large, requiring a 3.9 m turbine tip diameter for potassium and 2.8 m for
rubidium, for a 50 MW turbine [220]. Latent metal systems can also promote isothermal operation
requirements for many thermodynamic cycles including Stirling engines [224]. To achieve high
operating conditions, alkali metals [225], have high thermal conductivities that improve heat

88
transfer and allow a receiver to operate at high heat flux levels, while maintaining an acceptable
temperature difference between the absorber inner surface and the fluid [226].

4.4.1 Potassium and NaK

A research program for the development of a Rankine cycle based nuclear power plant for space
application in the 1960s demonstrated that potassium is a much better high temperature working
fluid compared to mercury. It is less corrosive to stainless steel at 870 °C than mercury at 480 ᵒC
or steam at 540 ᵒC. Additionally, potassium stainless steel boilers were operated from 5000 to
16,000 hrs. at 815–870 ᵒC with virtually no corrosion [227, 228]. Furthermore, a niobium system
worked for 10,000 hrs. at 1100 ᵒC with no trace of corrosion [228]. On the whole, potassium
boiling systems were operated for more than 200,000 hrs. at temperatures above 815 °C in the
development of long range nuclear electric space power plants [228]. Potassium compatibility with
stainless steel systems at a temperature much higher than that of standard steam cycles is due to
the favorable corrosion characteristics and to the low saturation pressure which implies minimal
stresses in heat exchanger tubes [220]. In Russia in 1972, a vapor potassium system was built and
successfully operated for more than 20 years. Evaporators, turbines, condensers, pumps and valves
were tested for around 10,000 hrs. at vapor temperatures up to 1030 ᵒC [220]. A temperature of
930 ᵒC was found to be compatible with the use of chromium–nickel stainless steel [227]. The use
of refractory metals (e.g. niobium, molybdenum, tantalum) has become attractive for high
temperature applications with potassium, whereas steel corrosion seems to be due largely to the
solubility of the constituting elements in this element. In the case of refractory metals this solubility
is extremely low; however, these materials do have oxidation issues and can have high costs [220].
For NaK the eutectic sodium-potassium alloy NaK78 (22.2-77.8 mass percentage [197])
melts at ambient pressure at -12.6 ᵒC and boils at 785 ᵒC [229]. Despite the inferior heat transfer
characteristics of NaK78 as compared to sodium, the low solidification point makes it very
attractive for transient power plants, like CSP such as Nuclear. However, freezing issues in pipes,
vents and receivers are practically eliminated. Other non-eutectic NaK alloys could show more
favorable thermodynamic properties for plants, like higher densities, at the cost of high
solidification temperatures [197]. Diver et al. [230] presented the state of the art in parabolic dish
CSP systems at the beginning of the 1990s where the investigators focused on indirectly heated
Stirling engines with liquid metal HTF that are evaporated in the receiver and condense on the heat
exchanger to the engine.

4.4.2 Lithium

At the highest temperatures predicted for molten metal utility in CSP plants [220], the use of
lithium as a heat transfer fluid could be attractive, owing to its minimal vapor pressure (normal
boiling point at 1310 ᵒC). Extensive tests on its corrosion characteristics at moderate temperatures
were performed worldwide owing to its essential role in nuclear fusion future reactors [220].
Unfortunately, its interaction with high alloy steels seems much worse than that of sodium (same
corrosion rate for lithium at 450 ᵒC and sodium at 700 ᵒC, [231]). Refractory metals, on the
contrary, seem compatible with lithium even at very high temperatures [220]. Additionally, lithium
has previously been proposed as heat transfer and heat storage medium at 1230 ᵒC top temperature

89
using a molybdenum alloy as building material [232]. In an investigation by Davis and Kikin,
[233], a 30 kW Rankine cycle system was operated for more than a 1000 hrs. at the Jet Propulsion
Laboratory (JPL) featuring a lithium heat carrier loop and a potassium power plant at a lithium top
temperature of 1150 ᵒC. Refractory metals were extensively used as a material for the hottest
components.

4.4.3 Sodium

Sodium is a capable HTF option for advanced high-temperature CSP receivers due to its high
operating temperatures and thermal conductivity, which results in high heat-transfer coefficients
and minimized heat-transfer surfaces. Sodium is relatively inexpensive and has a number of
distinct advantages over oils and salts in CSP applications, such as: raising the maximum receiver
temperature, increasing the maximum thermal flux of the receiver, lowering the operational
pressure of the primary HTF loop, and preventing solidification in the receiver piping [226]. HTFs
such as nitrate salt mixtures currently used in central receiver plants are capable of achieving bulk
temperatures of 600 °C, which is not suited for higher-efficiency power cycles such as ultra-
supercritical steam power cycles (~590-620 °C) [224] or advanced power cycles under
development (e.g. sCO2 Brayton cycle, 700 °C+) [64]. Previous sensible receiver systems using
sodium as a HTF have demonstrated enhanced receiver efficiencies between 90-96% ±10% [226]
at power levels up to 2.85 MWth, as well as systems with thermal efficiencies as high as 76% based
on daily energy averages [234]. For sensible sodium systems, it has been shown that external
receiver systems perform significantly better than cavity receivers due to reduced convection
losses and improved heat distributions [235]. However, receivers with latent sodium for Dish-
Stirling systems [237] have demonstrated evaporation operating temperatures between 700 °C-
850 °C. For a Dish-Stirling engine with a heat pipe containing latent sodium, efficiencies as high
as 93% were found for a peak absorber surface temperature of 830 ᵒC [219], where a 20%
performance improvement was realized with a directly illuminated receiver [226].
Latent sodium, being electrically conductive, can also be pumped by electromagnetic
pumps, thereby reducing the need for mechanical systems susceptible to operational mechanical
fatigue [226]. Sodium also does not need to be pressurized since its boiling point is much higher
than the receiver system’s operating temperature. Utility of latent sodium can achieve higher
receiver efficiencies and also reduce fatigue found with sensible systems. Although sodium pool
boiling has been studied extensively in both CSP and nuclear applications [64], forced convection
boiling has been studied far less, and has improved potential safety, principally due to lower latent
transport pumped mass. Latent systems exhibit lower pressures than pumped sensible systems,
where at 800°C system pressures are half an atmosphere and any failures result in air ingress rather
than sodium egress. Coupling a phase change storage material in the cycle minimizes hazards by
eliminating the need to have sodium and water in close proximity [226]. Latent sodium under a
low oxygen state is also compatible with many structural materials and does not have corrosion
issues (if O2 is properly controlled [226]) as found with sensible and other high-temperature HTF’s,
including molten salts. Boerema et al. [222] conducted a comparison of HITEC® Heat Transfer
Salt and sodium as HTFs for central receivers, where the main advantages for sodum, as illustrated
in Fig. 4.7, were higher operating temperatures, with lower pipe temperatures due to the high
thermal conductivity. This enables higher radiation fluxes, smaller aperture areas and subsequent
heat losses [222].

90
Figure 4.7. HITEC® heat Transfer Salt and sodium receiver efficiencies and maximum surface
temperatures for various pipe lengths [222].

4.4.3.1 Sodium Safety Issues

Sodium reacts with air and is inherently a potent reducing reagent and reacts violently with water
to form hydrogen and sodium hydroxide. It normally has an oxidation state of +1, and its single
valence electron is lost with great ease, yielding a colorless sodium ion, Na+ [226]. It reacts
violently with mineral acids, halogens, and reacts exothermically with oxidizing agents, organic
and inorganic halides, and protic media [229]. Sodium also reacts to generate shock-sensitive
products with sulfur oxides and phosphorous, and reacts with incandescence and many metal
oxides such as mercurous and lead oxides. Sodium dissolves in many other metals such as
potassium with great evolution of heat. The reactivity of a sample of sodium is largely related to
its surface area. Therefore, reactions involving solid pieces of sodium, especially with an oxide or
hydroxide coating, may be slow and controlled, whereas reactions with high-surface area sodium
dispersions may be vigorous [237].

4.4.3.2 Reactivity with Liquid Water

Contact of sodium with water or chlorinated hydrocarbons causes a violent exothermic reaction
and may cause detonation of the released hydrogen. Although sodium has a high chemical
reactivity with water, the heat release rate and heat of combustion is lower compared to
conventional hydrocarbon fuels [237]. However, in contact with water sodium often ignites the
hydrogen formed. Pure sodium begins to absorb hydrogen appreciably at approximately 100ᵒC
[238], where the rate of absorption increases with temperature. The reaction between sodium and
water can be divided into two half-reactions, where Eqn. 4.10 describes the loss of electrons from
sodium atoms, and Eqn. 4.11 describes when water molecules gain those electrons.

2[𝑁𝑎 → 𝑁𝑎 + + 𝑒 − ] (4.10)

2𝐻2 𝑂 + 2𝑒 − → 𝐻2 + 2𝑂𝐻 − (4.11)

91
The balanced equation for this reaction can then be written as:

2𝑁𝑎(𝑠) + 2𝐻2 𝑂(𝑙) → 2𝑁𝑎 + (𝑎𝑞) + 2𝑂𝐻 − (𝑎𝑞) + 𝐻2 (𝑔) (4.12)

For the reaction of water with liquid sodium, the ease of surface area expansion can produce more
explosive results since the reaction can be highly exothermic; where for a reaction at 25 ᵒC, where
sodium is a solid, approximately only 33.67 kilocalories per mole of heat is released [238].
Additionally, secondary reactions accompany violent reactions with water, with the liberation of
hydrogen gas which can be explosive. The heat liberated in Eqn. 4.13, is sufficient to cause the
reaction in Eqn. 4.14.
1
𝑁𝑎 + 𝐻2 𝑂 → 𝑁𝑎𝑂𝐻 + 2 𝐻2 + 𝑄̇𝐻𝑒𝑎𝑡 (4.13)

2𝐻2+𝑂2→2𝐻2𝑂 (4.14)

4.4.3.3 Reactivity with Air and Water Vapor

When exposed to air, sodium may ignite spontaneously at a temperature as low as 115 ᵒC,
depending on such conditions as humidity, dispersion, etc. For water vapor, sodium begins to
absorb hydrogen appreciably at approximately 100 ᵒC where the rate of absorption increases with
temperature [238]. A sodium fire can be recognized by a very low flame with many bright yellow
nodes. Sodium fires produce a dense white caustic smoke that contains highly alkaline material,
sodium monoxide (Na2O) and sodium hydroxide (NaOH), which can cause irritation and rapid
tissue destruction through chemical and thermal burns. As the temperature increases above 650
ᵒC, sodium can begin to spall concrete (which can cause secondary hazards such as flying debris),
consume asbestos, firebrick, transite and other similar materials [237]. Under high pressure
conditions, and temperatures above 300ᵒC, the formation of sodium peroxide is facilitated as the
very reactive sodium material is consumed before each O2 molecule can combine with enough
sodium to from Na2O, Eqn. 4.15.
2Na + O2 → Na2O2 (4.15)

Data on the combustion and burning rate of sodium in air and reduced oxygen atmospheres are
required to predict the rate of energy release for a fire of a given size. Conversely, when sodium
peroxide or monoxide are dissolved in water, oxygen is evolved and an alkaline solution is then
formed containing sodium hydroxide (Eqns. 4.16 and 4.17).
1
Peroxide: 𝑁𝑎2 𝑂2 + 𝐻2 𝑂 → 2𝑁𝑎𝑂𝐻 + 2 𝑂2 + 𝑄̇𝐻𝑒𝑎𝑡 (4.16)

Monoxide: 𝑁𝑎2 𝑂 + 𝐻2 𝑂 → 2𝑁𝑎𝑂𝐻 + 𝑄̇𝐻𝑒𝑎𝑡 (4.17)

During fires, a CO2-contianing atmosphere can arise that can transform sodium hydroxide
particles/droplets into solid sodium carbonate particles at a rapid rate, and as a function of relative
humidity (RH). As illustrated by Jordan et al. [239], this conversion at 50% RH occurs 5 times
faster than at RH levels less than 3% [239]. This may be explained that as the surface of wet
particles are already converted to solid crystal carbonate, it may act as a diffusion barrier, where

92
smaller particles have a higher content of sodium carbonate than larger ones. As shown in Fig. 4.8,
after 260 sec., only at RH levels above 50% is 100% sodium carbonate realized [239], which would
facilitate a safer environment after a fire since all volatile sodium aerosols will have progressed
fully through reaction.

Figure 4.8. Carbonate formation from sodium fires under varying relative humidity [239].
Through implementation of handling and engineering controls [31, 226], safe and optimal heat
transfer performance of CSP receivers using sodium can be realized. Solutions to address safety
can be broadly categorized according to prevention and extinguishment. Design controls of
receiver piping, such as utility of high conduction metals and absorber coatings, [64] have been
shown to reduce thermal stress, while design optimization [240] can be leveraged to reduce cycle
fatigue from induced vibration, which occurs in the direction of sodium flow. Flow boiling
instabilities can be abated through addition of small amounts of heavy inert gas (e.g. Xenon or
Argon) which provides pre-existing bubbles that can be inflated with minor superheat that
consistently provides low superheat startup and more stable operation [226].

4.4.4 Lead Bismuth Eutectic (LBE)

The lead bismuth eutectic composition Pb-Bi (44.5%-55.5%) has been investigated for
applications in CSP [241]. LBE alloy is a reasonably standard industrial HTF with a low melting
temperature (~124.6 ᵒC at 0.1MPa), though high boiling temperature (~1669.9 ᵒC), with low safety
risks [242]. The density of LBE changes little upon melting/solidification, however after
solidification, it is known to show a notable and time dependent volume evolution which could
cause unexpected material breakage. Prior research has provided the following operating
correlation recommendations for LBE [243], Table 4.3.

93
Table 4.3. Recommended thermophysical property correlations for LBE at 0.1 MPa and a Pb
44.5%, Bi 55.5% composition [243].
Temperature
Properties Correlation Range (K)
Density Ρ(kg/m3) = 11,096-1.3236T(K) 403-1,300
Dynamic Viscosity Μ(Pa-s) = 4.94x10-4exp(754.1/T(K)) 400-1,100
-2 -6 2
Heat Capacity Cp(J/kg-K) = 159-2.72x10 T(K) + 7.12x10 T(K) 400-1,100
Thermal k(W/m-K) = 3.61+1.517x10-2T(K)-1.741x10-6T(K)2 403-1,100
Conductivity
Sat. Vapor Pressure P(Pa) = 1.11x1010exp(-22,552/T(K)) 508-1,943

Like most molten metal HTF systems, temperature measurements are performed indirectly by
welding a thermocouple onto the outer surface of piping and heat exchange equipment. Since LBE
is a good conducting material, the error of these measurements has been generally approximated
to be negligible for relatively thin walls (approximately <1 in) [242]. Additionally, for LBE
applications, cover gas is generally recommended to displace oxygen down to levels less than
0.025%, where operating processes for LBE in CSP systems include: 1. Use of trace heaters to be
used to prevent solidification of LBE during pumping and transport and 2. Storage tanks and heat
exchange equipment charged with cover gas, such as argon [243].
According to Pacio and Wetzel [244], LBE has problematic drawbacks when compared to
sodium or NaK (reaction with water), and has a very high density, leading to much lower flow
speed requirements. The boiling temperature of LBE (1670 ᵒC) is even higher than that of sodium
but the solidification temperature (125 ᵒC) is higher as well, so that freezing is an issue. In a study
by Kim et al. [243] LBE was utilized in a developed test system where approximately 1000 kg of
LBE was successfully used to capture 35 MJ of solar energy. The test successfully transferred heat
from the solar air receiver to the LBE, with the temperature of stored LBE reaching 770 ᵒC.

4.5 Gases

The upper temperatures of gaseous HTFs in CSP systems are usually only limited by the materials
of the receiver pipes, ducts, etc. They are therefore especially suited for high-temperature
applications. Gases, however, have the downside of comparable low heat transfer coefficients and
densities. Gases can be used to directly power a gas turbine, thus making use of the very high
temperatures which can be generated in a central receiver. The exhaust gases can also be used to
power a bottoming cycle, which renders possible high thermal system efficiencies [197].

4.5.1 Air

Air is by far the most investigated gaseous HTF used in thermodynamic cycles as it is abundant,
with extensive experience with its utility as a heat transfer working fluid. It can be readily used in
a Brayton cycle, is non-hazardous, theoretically free of cost and does not necessitate a heat
exchanger for co-firing. The high operating temperatures of the fluid can enable combined cycle
plants with high efficiencies, however its heat transfer can be problematic due to its low density
and thermal conductivity. Therefore, it has unfavorable heat transfer characteristics, requiring
large heat exchangers and receivers for its application. For the implementation of a CSP receiver

94
in a gas turbine, it has to be pressurized, where performance is highly dependent on the pressure
drop between the compressor and expander [197].
Air can be used as an HTF as well as the power block working fluid where it can also drive
a Rankine steam cycle through a heat exchanger. For this application, an open air receiver is
typically used for simplicity reasons. The air doesn’t have to be pressurized (except to overcome
pressure drops in the receiver, piping and heat exchanger) and the turbine is not directly coupled
with the receiver outlet flow. However, the advantage of using a high temperature working fluid
cannot be made use of, since Rankine cycles are typically limited to approximately 640 ᵒC [197].
The only commissioned demonstration plant using this technology for grid-power is the Jülich
Power Tower, as described by Hennecke et al. [245]. In the AlSol project, this technology is also
used in a 7.1 MWe hybridized solar-natural gas power tower [246]. Additionally, Wilson
Solarpower [247] proposed a system in which unpressurized air is used as the HTF in a closed
receiver. This hot air then heats pressurized air in a regenerative heat exchanger, Fig. 4.9, to be
used in the microturbine within a Brayton cycle.

Figure 4.9. Regenerative air-air heat exchanger [247].

4.5.2 Helium

Helium possesses good heat transfer characteristics which could be beneficial in the design of a
solar receiver or combined power block cycle. Angelino and Invernizzi [220] suggested that a
closed helium cycle could be useful in cogeneration applications, as rejected heat is typically
available at relatively high temperatures. Based on system design parameters from a solar
application by Frutschi [248], the authors found that the Helium cycle was able to achieve an 800
ᵒC top temperature with a computed overall efficiency of 40.2%. Results of their analysis can be
seen in Table 4.4, where they also computed a potential top pressure for this system to be
approximately 20-60 bar. They concluded that the closed helium cycle lent itself intrinsically to
cogeneration where the heat of cooling of the working fluid prior to compression is available at a
comparatively high temperature [220].

95
 Table 4.4. Solar closed helium Cycle Study Results by Angelino and Invernizzi [220].

4.5.3 Water/Steam

Steam and supercritical steam (sH2O, Pcrit,H2O = 221 bar) has traditionally been used as a working
fluid in conventional power plants. For molten salt power towers that produce supercritical steam
for the power block, the efficiencies have seen a 2-3% improvement over superheated water
systems [203] since they have the potential to operate at higher temperatures and pressures [249,
250]. Operation with steam above the supercritical point can allow for higher temperatures and
fluid behavior that shares the compressibility of a liquid while avoiding complications due to
multi-phase heat exchangers, such as intrinsic instabilities. Supercritical water was an early
candidate for supercritical power cycles due to extensive operating knowledge of water with
Rankine cycles [241]. Ultra-supercritical (USC) steam plants have found high performance
conditions as high as 620 ᵒC / 310 bar [252]. USC turbines have found utility in coal plants where
their ratings can range between 200 MWe and 1050 MWe and reach net efficiencies in excess of
45%. Next generation advanced USC (A-USC) plants are expected to run at turbine inlet
temperatures of between 700 ᵒC and 760 ᵒC, with efficiencies above 50% [253].
Due to the high critical pressure of water, sH2O can require high cost materials for piping,
turbomachinery and heat exchangers. Additionally, it has been found that supercritical water is
extremely corrosive, not allowing its utility for use with turbines, compressors and recuperators in
Brayton cycles. Coventry and Pye [254] proposed a parabolic dish system, employing sH2O or
superheated (subcritical) water as the HTF and working fluid, with a molten salt storage system.
The advantage of sH2O systems is the lack of evaporation processes which lead to smoother
isobaric heating processes as compared to other superheated fluids [197]. The heat transfer
between sH2O and another single phase fluid, such as molten salt, is exegetically preferable as
demonstrated by Fig. 4.10 [254].

96
 Figure 4.10. Subcritical and supercritical heat transfer in systems with one HTF [254].

There has however been fouling concerns expressed as steam temperatures exceed 627 ᵒC due to
limitations of standard ferritic steels, where the requirement of more expensive high-nickel alloy
materials has been suggested [203].

4.5.4 CO2

Compared to water, carbon dioxide is less corrosive and has a much lower critical pressure of 73
bar, yet supercritical CO2 (sCO2) is relatively dense at approximately 0.6 kg/m3 [257]. These
characteristics can decrease stress on pipes, as well as flow speeds, which can reduce the size of
the power block turbomachinery and heat exchangers compared to steam. Additionally, sCO2
cycles show higher efficiencies than current [255] steam or air cycles. Other benefits of its higher
density is smaller machinery geometries and simpler plant design [258], where these benefits have
shown demonstrated promise in next generation nuclear [251] and conventional power plants, as
well as for cooling applications [256].
Although disadvantages may exist as outlined by Turchi [258] for sCO2 as an HTF and a
working fluid in CSP plants, Gary [259] suggested advantages for using sCO2 in combined cycle
solar power tower systems as a means of reducing the LCOE to 0.06 USD/kWhe. Chapman and
Arias [260] compared three parabolic trough configurations with synthetic oil, subcritical CO2 and
sCO2 as HTFs, where they found higher magnitude pumping power requirements necessary for
subcritical loops as compared to synthetic oil. However, due to the higher density of sCO2,
pumping power was significantly reduced even below that of synthetic oil. Though the high
pressure dramatically increased the requirements on pipes, where its utility will most likely not be
feasible in line-focus systems, but only in central receiver systems [197].
Chapman and Arias [260] as well as Ma and Turchi [261] addressed the problem of adding
a thermal energy storage system (TESS) to a sCO2 system, where a molten salt storage medium
was found to limit the turbine inlet temperature during discharging processes to temperatures less
than 600 °C. Active direct and passive storage systems however do not appear viable due to the
need for high pressure tanks. Cox [262] found that standard heat transfer correlations are in general
applicable to sCO2 heat exchangers, though in close proximity to the critical point the measured
heat transfer noticeably differed from the calculated values.

97
Rouillard et al. [263] compared the corrosion effects on a ferritic-martensitic and different
austenitic steels in contact with sCO2 at a temperature of 550 ᵒC and a pressure cycle of 250 bar.
Their research was aimed at the development of components for the secondary cycle for advanced
nuclear reactors. It turned out that corrosion had a much more profound effect on ferritic-
martensitc steels as compared to austenitic ones [197]. Dostal et al. [264] gave an elaborate
overview on sSO2 as a working fluid and HTF in nuclear power plants [197]. The information
found in their report is applicable to CSP, as they investigated different configurations for liquid
metal cooled reactors with sCO2 as the working fluid. One safety advantage realized is that there
is no direct heat exchanger between a potentially volatile liquid metal and a water/steam cycle
[197]. CO2 also reacts exothermally with liquid metals, such as sodium, but no hydrogen is created
in the reaction, therefore greatly decreasing the hazards associated with liquid metals, which have
the capability of realizing high temperatures and subsequent efficiencies [224]. Dostal et al. [264]
built a model for different sCO2 cycles for preheating, reheating, precooling or intercooling
applications, and compared them to a supercritical water and ideal gas Brayton cycle with helium
as the working fluid. In this analysis they investigated the influence of heat exchanger (pre-cooler,
regenerator, etc.) sizes on thermal efficiency and pressure losses. The investigators found great
potential for sCO2 cycles due to the simple layout of Brayton cycles as compared to Rankine
cycles, and high efficiencies at moderate pressures/temperatures [197]. The latter caused by the
cycles’ low compression work due to the low compressibility factor of sCO2 at the critical point.

98
5. THERMAL ENERGY STORAGE
5.1 Overview

The extended operation of solar thermal power plants using stored thermal energy can significantly
improve the power dispatch ability [267]. Thermal energy storage systems use materials that can
be kept at high temperatures in insulated containers. The heat retrieved can then be used in
conventional thermal power plants for power generation at times when sunlight is not available or
when weather conditions are not favorable [266]. Thermal energy storage can be less expensive
overall than electric batteries and capacitors, which have many reliability issues [293]. The
reduction in costs for thermal energy storage can allow CSP plants to increase the capacity factor
and dispatch power as needed. Although direct energy storage in batteries and capacitors may have
a high efficiency, it is still very challenging and expensive, particularly when storing a large
quantity of electrical energy [265]. Storage provides heat necessary for operation of thermal power
plants when sunlight availability is out of phase, and thus increases their operational capacity (in
terms of the daily operational time).
From an energy efficiency perspective, direct use of solar thermal energy for heating is
much more efficient than using electricity for heating, as electricity generation requires much more
input of other types of energy than the generated electricity [266]. Financially, TES reduces
mismatch between solar energy supply and utility demand [267], where TES systems can be
integrated within CSP systems to facilitate buffering during transient shading events (clouds, dust,
etc.), where the power block of a CSP facility can be severely affected that can reduce turbine
efficiency and contribute to start-up losses [268]. Although many HTFs, and even the solid walls
of heat exchange equipment can retain some thermal inertia during operation [269], at a utility-
scale, turbine shut down can still become imminent under shading conditions [270]. TES has the
ability to dramatically improve this buffering to not only improve the performance of CSP systems,
by allowing for continued power production, it can also impact the reliability of power block
components by reducing thermal-mechanical stresses induced by system losses [268].
Table 5.1 shows the capacity-credit results for a subset of solar multiples and thermal
storage capacities [271]. Configurations with less than six hours of energy storage capacity would
result in lower capacity credit [272]. The flexibility of CSP should also allow it to meet “flexible
capacity” requirements, such as the California Public Utility Commission (CPUC) requirement
that all load-serving entities procure capacity with sufficient flexibility to address the largest
predicted 3-hour ramp rate in each month [273].

Table 5.1. Capacity credit (%) for various configurations of CSP-TES [265].
Solar Multiple
TES
0.7 1.0 1.3 1.7 2.0 2.5 3.0
(hours)
6 93% 96% 97% 98% 98% – –
9 – – – 98% 98% 99% –
12 – – – – – 99% –
15 – – – – – 99% 99%
18 – – – – – – 99%

99
Fig. 5.2 describes the result of a recent NREL study comparing the combined operational and
capacity benefits of CSP-TES relative to PV under varying levels of renewable penetration in
California [265]. The analysis found that the value of CSP, compared to variable-generation PV,
demonstrated an increase in value of up to 6 cents/kWh under a 40% RPS [274].

Figure 5.2. Total value, which includes operational and capacity value, of CSP-TES and PV under
33% and 40% RPS scenarios [274].

The necessity of thermal storage purpose is twofold: to increase the value of the power generated
by strongly reducing its aleatory character and to improve the plant economics by using the
available hardware (structures, piping, power block, grid connections, etc.) for more hours a year
[220]. CSP with thermal energy storage (or CSP-TES) provides considerable flexibility, increasing
its own value to the grid and even enabling greater grid penetration of other variable-generation
technologies such as PV and wind [274]. Previous analysis has demonstrated the ability of CSP to
provide frequency-response reserves that can help maintain grid stability [275]. Another
significant benefit of CSP-TES is its ability to provide firm capacity, which is the capacity
available to the power system at times of greatest need, which is most often the hours of highest
net demand [272, 276]. Variable-generation resources such as PV and wind tend to be limited in
availability during these hours at high penetration of renewables on the grid because these hours
typically occur when the solar or wind resources are unavailable [277]. As a result, installation of
these technologies alone cannot meet system peak demand, especially as older fossil plants are
retired and new capacity additions become necessary to maintain system reliability [274].
According to Heller [197] there are two primary means for storing energy in CSP systems.
The active direct methodology is considered [200] the simplest way of storing heat, which uses
the primary HTF as the storage medium, Fig. 5.3a. This has been shown to work particularly well
when using synthetic oil [197] or molten salts [185] that remain in a liquid state at elevated
temperatures.

100
 a. b. c.
Figure 5.3. Types of thermal energy storage, a. active direct, b. active indirect and c. passive [197].

In DSG pants or when gas is the primary HTF, either a gaseous medium would have to be stored,
resulting in low volumetric energy capacities, or a high pressure needs to be maintained inside a
tank to condense the medium, reducing long-term storage viability [197]. In active indirect storage
systems (Fig. 5.3b), the storage medium is separated from the primary HTF by a heat exchanger,
which has the advantage of not requiring the HTF to serve as the heat storage medium, providing
more design flexibility.
Concentrating solar plant projects like Andasol in Spain [278] incorporate heat storage that
allows a system to operate for approximately 6 to 12 hrs. in the absence of incident solar irradiance,
where advanced TES technologies use a two-tank salt system, such that the hot and cold tanks
store liquid salts separately [197]. This system can be used since components associated with
molten-salt handling (pumps, valves, tanks, and heat exchangers) are available at commercial scale
for nitrate salts (NaNO3 and KNO3), within salt applicable temperatures [279]. In a two-tank CSP
system, salt flows from the cold tank to the solar receiver, is heated, and enters the hot tank. During
power generation, hot salt flows to a steam generator and returns to the cold tank [197].
Current research and industrial applications with CSP thermal energy storage can be
principally divided into three primary categories: 1. sensible, 2. latent and 3. chemical (sorption
and thermochemical), which are illustrated by Fig. 5.4. This figure by Pfleger et al. [442] provides
some examples of different forms of thermal storage media which will be discussed further in the
next subsequent sections.

Figure 5.4. Classification of CSP thermal storage media [442].

101
5.2 Sensible Energy Storage

“Sensible heat” indicates that the storage process can be sensed by a change in temperature, where
the relation of the change in temperature and the stored heat is directly related by the heat capacity
Cp. TES systems for CSP generally store energy by changing the temperature of the storage
medium, which is then retained for later utility in power production. In sensible TES systems, heat
energy is stored or released by heating/cooling an HTM [139]. The amount of energy input to a
TES in a sensible heat system is related to the mass of storage material and its heat capacity, as
well as the temperature difference of the storage medium between its initial and final states. This
heat transfer 𝑄̇𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 can be expressed as:

𝑄𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 = 𝑉𝜌𝐶𝑝 (𝑇𝑖+1 + 𝑇𝑖 ) = 𝑚𝐶𝑝 ∆𝑇 (5.1)

where V is the volume of the storage materials, m and Cp denote the mass and specific heat of the
storage material, and ΔT is the temperature difference before (at an arbitrary time i) and after the
storage operation (at an arbitrary time i+1). The specific heat of solids can be estimated by the
Dulong-Petit law:
𝑞𝐶𝑝,𝑠𝑜𝑙𝑖𝑑 ≅ 3𝑅Δ𝑇 (5.2)

3𝑅
𝐶𝑝 ≅ (5.3)
𝑀

where 𝑞𝐶𝑝,𝑠𝑜𝑙𝑖𝑑 is the heat stored per mole, R is the universal gas constant and M is the molar mass.
An investigation by Weinstein et al. [139] found that a mole of sensible TES material almost
always has the same heat stored for an identical Δ𝑇, which is approximately 25 J/(K-mol) or
equivalently as a heat capacity of 3kBT for each atom from the equipartition theorem, where kB is
the Boltzmann constant. The investigators also determined from Fig. 5.5 that the relative volume
of atoms in the lattice (packing fraction) does not change significantly between materials, which
leads to similar volumetric energy densities 𝜌𝐶𝑝 (<one order of magnitude).

Figure 5.5. Volumetric heat capacity of TES materials, where curve discontinuities correspond to
phase change [139].

Energy storage in sensible TES systems are determined largely by the volume of the system where
total capacity of the system (kWht/m3) can be determined by integration of the volumetric heat

102
capacity over the operating temperature range [139]. Values for some sensible TES materials with
these volumetric heat capacity values are shown in Table 5.2 provided by Weinstein et al. [139].

Table 5.2. Sensible CSP TES materials, where material properties are representative of the
specified operating range [139].

Sensible systems can also be comprised of solid particles, such as those used in falling particle
receivers. Dense gas-particle suspensions of approximately 50% of the solid, have been proposed
for application in direct and indirect tubes as a HTF. This relatively new HTF behaves like a liquid
with extended working temperatures as it remains in a fluidized state at any temperature above
freezing and permits working temperatures up to 700 °C or more. Moreover, it can be used as an
energy storage medium because of its high thermal capacity and potential ability to be employed
in heat exchangers for power block dispatchable energy generation. The addition of nanoparticles
in fluidized HTFs have also been explored to facilitate nanofluidic transport, which has been found
to further enhance physical and transport properties, where work by Nagarajan et al. demonstrated
a collector efficiency increase as high as 23.3% when using an Al2O3/water nanofluid, with a
weight concentration of 0.2%. [446].

5.2.1 Molten Salts

The use of molten salts in a sensible storage medium is the standard against which most current
storage options are compared. Conventional molten nitrate/nitrate salts become unstable at
approximately 600 °C, however freeze at approximately 228 °C [7], which can be favorable for
high-efficiency CSP receivers. Achieving higher-temperature, higher-efficiency power cycles
requires new higher-temperature compositions [7]. However, many molten salts solidify at
temperatures above ambient, causing blockage and potential damage to piping and heat
exchangers. To address this, lower freezing-point salts would aid in the mitigation of freezing
events in addition to achieving much higher temperatures [280]. Carbonate salts also appear to be

103
suitable for operation up to 850 ᵒC [281]. Further work by Fraas [228] is described for a molten
salt cooled nuclear reactor with a top temperature of 980 ᵒC. Research with chloride salts have
also been investigated as a thermodynamically stable means for achieving higher CSP operating
temperatures (>700 °C) [7]. Sand, sintered bauxite, alumina, and ceramics have also been
investigated as a heat-transfer and storage media [282]. Desirable features include a high solar
absorptance (for direct irradiated particles), low thermal emittance, good durability, resistance to
sintering, high heat capacitance, ease of flow, and low cost. Studies [7] have shown that
commercial ceramics particles (sintered bauxite) used as proppants for hydraulic fracturing have
many of the desired properties for directly irradiated receivers.
Additionally, solid-state storage systems have recently been researched [283] due to the
relative low cost of many potential solid-state storage media and their ability to achieve higher-
temperature operation. However, heat transfer into and out of the solid-state materials represents
the largest challenge in developing these systems.

5.3 Latent Thermal Energy Storage

Most thermal energy storage systems in operation are sensible; however, storing heat in the form
of latent heat of PCMs can significantly increase energy density [290]. Latent thermal energy
storage (LTES) is facilitated by converting a storage media from one phase to another within an
isothermal process. This can provide enhanced storage capacity when compared to sensible heat
thermal energy storage (STES) systems of the same temperature range since STES materials are limited
by the need to store at higher temperatures to achieve the desired output temperature for turbine
systems [456]. The latent heat of phase change can store or retrieve large amounts of energy and
are generally smaller, and on a weight basis LTES materials, have the potential to be more effective
than sensible heat storage materials. LTES commonly leverages phase change phenomena between
solid and liquid states, due in large part to its lower volumetric expansion, compared to liquid-gas,
and its high latent heat compared to solid-solid transition [456]. In latent TES systems, energy is
stored during phase change, and due to the specific heat of a typical medium the latent enthalpy
change is usually greater than the sensible heat change for a given system size [284].

Figure 5.6. Volumetric latent heat with respect to melting temperature for latent heat TES PCM
materials [139].

104
The U.S. DOE requirements for TES under the SunShot Initiative seeks to improve heat transfer
and thermal energy storage medium, lower the systems costs to less than 15 kW/h, increase the
exegetic efficiency to greater than 95%, and lower material degradation due to corrosion to less
than 15 m/year [457]. A variety of LTES have been proposed to achieve desired CSP efficiencies
and energy capacities, where in general these systems are categorized in two groups. First, active
systems, where the storage medium is a fluid and flows between the tanks (common in sensible
heat thermal energy storage). The second or passive systems, is where the storage medium is solid,
contained in one tank, and the heat transfer fluid passes through the storage material only for
charging and discharging [456]. However, most LTES systems are passive to ensure phase change
within a controlled environment.
As can be noted from the boiling curve Fig. 2.3 in section 2, latent heat storage materials
are usually useful over a small temperature range. The stored energy during a latent storage process
can be assessed as:
𝑄𝐿𝑎𝑡𝑒𝑛𝑡 = 𝑉𝜌Δℎ𝑚 = 𝑚Δℎ𝑚 (5.4)

where m denotes the mass and Δℎ𝑚 is the specific latent enthalpy difference (heat of fusion) of the
PCM material. Weinstein et al. [139] suggested that latent heat of melting roughly scales linearly
with melting temperature for various classes of materials, where higher energy densities can
typically be achieved for PCMs with high melting temperatures, Fig. 5.6. The investigators also
pointed the advantages of latent heat PCMS over sensible materials where the energy density is
typically higher as ℎ𝑚 ≫ 𝐶𝑝 Δ𝑇 and that charging and discharging processes usually occur at
constant temperature, which is advantageous for alignment with system design heat engine
operating temperatures [139].
An investigation by Angelino and Invernizzi [220] demonstrated that cooling of a molten
salt from 550 ᵒC to 300 ᵒC facilitates an available specific energy of 387 kJ/kg while solidification
of LiF at 848 ᵒC supplies a specific energy of 1087 kJ/kg. A large number of pure salts or mixtures
(typically fluorides of Ca,Mg, Na, K, etc.) are available with melting temperatures ranging from
750 to 1250 ᵒC [220, 285]. Technical challenges still need to be addressed for the large volume
change during melting with low thermal conductivity of the liquid and of the solid, where solutions
have been proposed [286, 287]. A wide preliminary screening [288], where the fluoride salt is
completely free from water, demonstrated that many metals and alloys are available with a
sufficient corrosion resistance at temperatures around 900 ᵒC. In particular, mild steel and pure
nickel exhibits a relatively good resistance [285]. At high temperatures for binary cycles, phase-
change heat storage seems the most promising option, where the thermal characteristics of melting
PCMs, that exchange heat at a fixed temperature, simplifies insertion of the heat storage in the
plant layout [220]. An instance of CSP plant layouts which includes a phase change thermal storage
is given by Wallin and Dustin [286].
However, the process is isothermal, so unless the power cycle relies on nearly isothermal
conditions (e.g., Stirling cycle, heavily recuperated sCO2 cycles), high exergetic losses may occur.
It has been demonstrated that TES systems that employ a latent phase change material (PCM) can
maximize storage of useful energy (or exergy) and at the system level, increase the efficiency of a
power system [289]. Nithyanandam et al. [291] illustrated that the energy required to melt one
kilogram of sodium nitrate by latent heating, is 75% higher compared to the energy required to
raise the temperature of one kilogram of the same substance by 1 K through sensible heating. Thus
TES systems using latent sodium-based phase change materials have the advantage of being

105
volumetrically compact in size with enhanced heat transfer coefficients. Thermodynamically,
sensible systems have been shown to be a poor exergetic fit for high temperature CSP systems,
which include highly recuperated sCO2, phase change storage, hybrid storage and chemical
processing [7].
Finally, advanced research of various encapsulation methods for PCM have been evaluated
to address challenges with heat transfer while the system is discharging [17]. Examples include
encapsulation of small amounts of PCM within a solid that will not undergo phase change and
impregnation of PCM into metal or graphite foams. The use of PCMs with an embedded
thermosyphon or heat pipes have also been investigated to reduce thermal resistances [292].

5.4 Thermochemical Energy Storage

More compact TES can be achieved based on energy storage designs that utilize chemical
reactions. Thermochemical storage (TCS) is based on the heat capacity and its change as function of
temperature accompanied with chemical reaction. Here, TCS makes use of the enthalpy of reaction
ΔH where reactions featuring a positive change of ΔH (endothermic reaction) heat can be stored
[284]. Conversely, energy can also be released by a backward reaction (ΔH < 0) which can provide
energy input to a power block. This class of energy storage includes sorption and thermochemical
reactions. In thermochemical energy storage, energy is stored after a dissociation reaction and then
recovered in a chemically reverse reaction. Thermochemical energy storage has a higher storage
density than the other types of TES, allowing large quantities of energy to be stored using small
amounts of storage substances. Energy storage based on chemical reactions is particularly
appropriate for long-term storage applications, e.g., seasonal storage of solar heat, because the
process involves almost no energy losses during the storing period [284]. In their work, Abedin
and Rosen [284] performed a comparison between sensible, latent and thermochemical energy
storage where their results can be seen in Table 5.3, where the authors found that each had different
operating characteristics and distinct advantages and disadvantages.

The main principle of thermochemical TES is based on a semi-reversible reaction [284]:

𝐶 + ℎ𝑒𝑎𝑡 ⇆ 𝐴 + 𝐵 (5.5)

In Eqn. 5.5, a thermochemical material, C absorbs energy and is converted chemically into two
components, A and B, which can be stored separately. The reverse reaction occurs when materials
A and B are combined together and C is formed [284]. Energy is released during this reaction and
constitutes the recovered thermal energy from the TES material. Storage capacity of this system is
dependent on the heat of reaction when material C is formed. The charging process for this state
is endothermic where thermal energy is absorbed from an energy resource and used for dissociation
of the thermochemical material. This is equivalent to the heat of reaction or enthalpy of formation
[284]. After the charging process, components A and B are separately stored with little or no energy
losses where the materials are stored at ambient temperatures, leading to no thermal losses (except
during the initial cooling of components A and B after charging) [284].

106
Table 5.3. General comparison between thermochemical energy storage with sensible and latent
systems [284].

Any other energy losses are due to degradation of the materials. Finally, during the discharge
process, constituents A and B are combined in an exothermic reaction. The energy released from
this reaction permits the stored energy to be recovered. After discharging, component C is
regenerated and can be used again in the cycle. The discharging reaction can be written as:

𝐴 + 𝐵 → 𝐶 + ℎ𝑒𝑎𝑡 (5.6)

Another thermochemical process, sorption (adsorption and absorption) is based on a chemical

processes and thus are also considered chemical heat storage. Adsorption occurs when an
adsorptive material accumulates on the surface of an adsorbent and shapes a molecular or atomic
layer [284]. The adsorptive material can be a liquid or gas while the adsorbent can be a solid or
liquid. Absorption is a process that occurs when a substance is distributed into a liquid or solid and
forms a solution. Storing solar energy in the form of chemical bonds is attractive because the
energy can be stored indefinitely, and thermal boosting can occur in the form of
reduction/oxidation reactions [284]. There has been much research in this area where
investigations by Weber and Dorer [294] analyzed long-term heat storage using a closed sorption
system with NaOH and water as the working pair and compared the results with a conventional
storage system, focusing on system volume [298]. An investigation based on bromide strontium
as the reactant and water as the working fluid, in a system using flat plate solar collectors and
applied to direct floor heating, demonstrated the relationship between attained power levels and
the heating storage capacities of reactive composites [284]. Mauran et al. [299] analyzed

107
experimentally the same working pair for heating and cooling purposes (heating in winter or mid-
season and cooling in summer). Zondag et al. [300] characterized magnesium sulphate as a storage
media and examined cycling behavior of MgSO4 and the dehydration temperature of the reactant.
Thermal energy storage based on a Ca(OH)2 and CaO cycle is also an example of thermochemical
energy storage, where the reversibility and efficiency of this system was investigated in Azpiazu
et al. [301]. Thermochemical energy storage based on ammonia and water has also been
investigated in conjunction with a solar thermal plant [284]. General characteristics of these
working fluis, as well as a dissociation and synthesis reactor were studied in [302], where the
optimization of a related heat recovery device has been reported [303]. Overall, kinetics, energy
density, and techno-economic costs present challenges and limitations to applying thermochemical
energy storage [284].

108
6. CSP THERMODYNAMIC POWER SYSTEMS

6.1 Overview

Solar energy has a high exergetic value as it originates from processes occurring at the sun’s
surface with a blackbody equivalent temperature of approximately 5777 K [187]. Therefore, more
than 93% of the energy generated may be theoretically converted to mechanical work by
thermodynamic cycles [304], or to Gibbs free energy by chemicals and solarized chemical
reactions [305], including promising hydrogen production processes [306]. Previous techno-
economic analyses have shown a significant potential for cost reduction through efficiency
improvement of the power block [66, 307]. Due to widespread use of turbomachinery and heat
engine technology, CSP has some unique challenges with respect to other renewable energy
technologies as a heat source in that it stands to benefit from economies of scale and
thermodynamic technological advances from industries such as coal, natural gas and nuclear
industries. While modern subcritical steam cycles (which are also common thermodynamic power
cycles to date for CSP) may be limited to thermal efficiencies up to approximately 42% [66],
supercritical steam cycles have been developed with thermal efficiencies exceeding 47% [249],
with its relative comparison to other thermodynamic cycles as presented in Fig. 6.1. Combined
cycles, which use rejected heat from a high-temperature cycle to drive lower-temperature cycles
to supplement the power output, typically offer higher thermal efficiencies (potentially exceeding
60%) and have been used in traditional power cycles for decades [308]. Such high-performance
cycles have been shown to be adaptable for CSP systems [309, 310], and represent an important
step in reducing LCOE, while promoting CSP as a true competitor to traditional methods for
utility-scale power generation. Currently, there is much effort to increase thermal conversion
efficiency to meet the U.S. department of Energy (DOE) SunShot Initiative target of >50% thermal
efficiency in an effort to reduce the cost of solar energy [7].

Figure 6.1. CSP thermodynamic cycle thermal conversion efficiency comparisons [311].

In order to achieve high conversion efficiencies, high temperatures at the heat engine and low
temperatures at its energy sink are necessary. These temperatures are limited by the HTF in the
receiver, as well as the working fluid in the heat engine cycle [197]. Overall, most ideal

109
thermodynamic cycles operate based on four processes of compression, heat input, expansion and
heat rejection. Table 6.1 provides a breakdown comparison between various cycles that CSP
currently, and potentially, can employ for research and application depending on whether the
respective system operates independently or in combination with an internal or external
combustion cycle.
Table 6.1. Thermodynamic Cycle Process Comparisons.

Single cycles represent the simplest thermal power plant configurations, and are traditionally the
baseline in identifying high-efficiency systems. A combined cycle however utilizes multiple
thermodynamic cycles: one primary high-temperature (topping cycle) and one or more lower-
temperature cycles (bottoming cycle), driven by heat rejected from the higher-temperature cycle.
According to Dunham and Iverson [203] although significant improvements have been made with
single cycle systems, it has been widely believed [312, 313] that operation above 50% thermal
efficiency requires implementation of combined cycles with traditional HTFs. However, it has also
been suggested that advanced sCO2 and helium Brayton cycles could potentially achieve 50%
[203]. For example, this can be attained by a Brayton-topped, combined cycle which can allow for
optimum use of highly concentrated solar energy where the ideal operating temperature is beyond
high-temperatures ever attained by sub-critical Rankine cycles [314]. While modern subcritical
steam cycles may be limited to thermal efficiencies up to approximately 42% [66], supercritical
steam cycles have been developed with thermal efficiencies exceeding 47% [249]. Combined
cycles, which use rejected heat from a high-temperature cycle to drive a lower-temperature cycle
to supplement the power output, typically offer higher thermal efficiencies (potentially exceeding
60%), and have been used in traditional power cycles for decades [308]. Such high performance
cycles have been shown, or considered to be adaptable to CSP systems [309, 310].

6.2 Carnot Cycle & Thermodynamic Limitations

When assessing power block cycles, it is important to have a relative fundamental thermodynamic
basis by which to compare them. The Carnot cycle provides an upper limit for thermodynamic
efficiency that any classical thermodynamic engine can achieve during the conversion of heat into
work, or conversely, the efficiency of a refrigeration system in creating a temperature difference
to facilitate work to the system. Every thermodynamic system exists in a particular state, where a
Carnot cycle exhibits the different states a system can go through, where it can eventually be

110
returned to its initial state. In the process of going through this cycle, the system may perform work
on its surroundings, thereby acting as a heat engine, Fig. 6.2.

Figure 6.2. Carnot cycle work and thermal power transfer between two heat reservoirs

The process states of a Carnot cycle as depicted by the PV and TS diagrams in Fig. 6.3, when
acting as a heat engine.

a. b.
Figure 6.3. Carnot cycle thermodynamic P-V and T-S diagrams.

Process 1-2: Reversible isothermal expansion of a gas at a temperature, TH (isothermal heat

addition or absorption). During this step the gas is allowed to expand, where it does work on the
surroundings. The temperature of the gas does not change during this process, and thus the
expansion is isothermal. The expansion of gas is facilitated by absorption of thermal energy 𝑄̇𝑖𝑛
and entropy 𝑑𝑆 = 𝑄̇𝑖𝑛 ⁄𝑇𝐻 from the high temperature reservoir.

Process 2-3: Isentropic (reversible adiabatic) expansion of the gas (isentropic work output). For
this step the mechanisms of the engine are assumed to be thermally insulated, thus they neither
gain nor lose heat. The gas continues to expand, doing work on the surroundings, and losing an
equivalent amount of internal energy. The gas expansion causes it to cool to the temperature TC.
For this process the system entropy remains unchanged.

111
Process 3-4: Reversible isothermal compression of the gas at temperature TC. (isothermal heat
rejection). Here the ambient surroundings do work on the working fluid facilitating heat energy
𝑄̇𝑜𝑢𝑡 and entropy 𝑑𝑆 = 𝑄̇𝑜𝑢𝑡 ⁄𝑇𝐶 to flow out of the working fluid to the low temperature reservoir.
For this theoretical cycle, this would be the same amount of entropy absorbed in process 1-2, as
can be seen from the Clausius inequality for an irreversible cyclic process:

𝛿𝑄̇
∫ ≤0 (6.1)
𝑇

Process 4-1: Isentropic compression of the working fluid (isentropic work input). Here, just as
with process 2-3, the mechanisms of the heat engine are assumed to be thermally insulated. During
this process the surroundings do work on the working fluid, increasing its internal energy and
compressing it, causing the temperature to rise to TH. The entropy for this process remains
unchanged.

From the Clausius inequality Eqn. 6.2, the amount of entropy S added to the system during the
cycle is defined as:
𝛿𝑄̇
∆𝑆 = ∫ (6.2)
𝑇

As prescribed by the 2nd law of thermodynamics, entropy is a state function since it depends only
upon the state of the system and not what path the system took to get there. This however is in
contrast to thermal energy added as heat (𝛿𝑄̇ ) and work (𝛿𝑊̇ ), which may vary depending on the
path. If the amount of energy added by heating can be measured during the process, and the
temperature can be measured during the process, the Clausius inequality can be used to determine
whether the process is reversible or irreversible by carrying out the integration in the Clausius
inequality. In a cyclic process, the entropy of the system at the beginning of the cycle must equal
the entropy at the end of the cycle. In the irreversible case, entropy will be created in the system,
and more entropy must be extracted (ΔS < 0) than was added in order to return the system to its
original state. In the reversible case, no entropy is created and the amount of entropy added is equal
to the amount extracted. However, for real systems, irreversibility I exists and accounts for the
amount of exergy destroyed, or work wasted in a closed system. Irreversibility as it relates to the
exergy in a closed system, with respect to entropy generated, is described by:

𝐼 = 𝑇0 𝑠𝑔𝑒𝑛 (6.3)

where I > 0 if irreversibilities are present in the system thereby defining the actual work as:

𝑊̇𝑎𝑐𝑡 = 𝑊̇𝑚𝑎𝑥 − 𝐼 (6.4)

With the relation 𝑊̇𝑚𝑎𝑥 = ∆𝐵 = 𝐵𝑖𝑛 − 𝐵𝑜𝑢𝑡 the available work or exergy is related to the
irreversibility of a closed, real system as:

∆𝐵 = 𝑊̇𝑎𝑐𝑡 − 𝐼 (6.5)

112
For a Carnot engine with an unlimited reservoir that remains unaltered by the system, the exergy
of a system heading towards equilibrium with respect to time is described by following
expressions:
𝑑𝐵
≤0 (6.6)
𝑑𝑡

𝑑𝑆𝑡𝑜𝑡
≥0 (6.7)
𝑑𝑡

From further integration, the area under the curve of Eqn. 6.8 can then be interpreted as the thermal
power transferred into or out of the system.
𝐵
𝑄̇ = ∫𝐻 𝑇𝑑𝑆 (6.8)

If a process moves to a state of greater entropy, the area under the curve will be the amount of heat
absorbed by the system, conversely if the process moves towards a state of reduced entropy it will
correspond to heat removal. The area inside the cycle will then be the difference between the two,
however since internal energy of the system must return to its initial value, the difference will
correspond to the amount of work done by the system over the cycle. These relationships can be
understood analytically by assessing a closed-system reversible process where the amount of work
done over a cyclic is:

𝑊̇ = ∫ 𝑃𝑑𝑉 = ∫(𝑑𝑄̇ − 𝑑𝑈) = ∫(𝑇𝑑𝑆 − 𝑑𝑈) = ∫ 𝑇𝑑𝑆 − ∫ 𝑑𝑈 = ∫ 𝑇𝑑𝑆 (6.9)

Here since dU is an exact differential, where its integration over a closed loop is zero, it follows
that the area inside the loop on a T-S diagram is equal to the total work performed if the loop is
traversed in a clockwise direction, and is equal to the total work done on the system if the loop is
traversed in a counterclockwise direction. Finally, the theoretical Carnot efficiency can be written
in terms of output work and heat input, as well as the volumes at different states by:

𝑊̇𝑜𝑢𝑡 𝑄̇ 𝑇 [𝑙𝑛(𝑉 ⁄𝑉 )]
𝜂𝐶𝑎𝑟𝑛𝑜𝑡 = = 1 − 𝑄̇ 𝐶 = 1 − 𝑇 𝐶 [𝑙𝑛(𝑉4 ⁄𝑉3 )] (6.10)
𝑄̇𝑖𝑛 𝐻 𝐻 2 1

The processes from states 2 to 3 and from 4 to 1 are both adiabatic and reversible where the
expression 𝑃𝑉 𝛾 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 holds. From the ideal gas equation of state, one is able to obtain
𝑇𝑉 𝛾 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡, which provides:
𝛾−1 𝛾−1 𝛾−1 𝛾−1
𝑇𝐻 𝑉2 = 𝑇𝐶 𝑉3 and 𝑇𝐻 𝑉1 = 𝑇𝐶 𝑉4 (6.11)

Therefore, with incorporation of these expressions into Eqn. 6.10:

𝑉 𝛾−1 (𝑇 ⁄𝑇 ) 𝑉 𝛾−1 𝑉4 𝑉
(𝑉4 ) = (𝑇𝐻⁄𝑇𝐶 ) (𝑉1 ) → = 𝑉1 (6.12)
3 𝐻 𝐶 3 𝑉3 3

By evaluation of Eqn. 6.10, one is able to observe that heat received or rejected is related to the
temperatures of the isothermal portions of the cycle by Eqn. 6.13, and a final form of Carnot
efficiency Eqn. 6.14:

113
 𝑄̇𝐿 𝑄̇
+ 𝑇𝐻 = 0 (6.13)
𝑇𝐿 𝐻

𝑇
𝜂𝐶𝑎𝑟𝑛𝑜𝑡 = 1 − 𝑇 𝐶 (6.14)
𝐻

For more realistic, less efficient cycles that contain irreversibilities, the average high and low cycle
temperatures 〈TH〉 and 〈TC〉 will be lower than TH and TC respectively.

1
〈𝑇𝐻 〉 = ∫ 𝑇𝑑𝑆 (6.15)
∆𝑆 𝑄̇𝑖𝑛

1
〈𝑇𝐶 〉 = ∫ 𝑇𝑑𝑆 (6.16)
∆𝑆 𝑄̇𝑜𝑢𝑡

6.3 Rankine Cycle

6.3.1 Overview and Theory

A single Rankine power cycle represents one of the simplest power block configuration, and can
be the baseline for identifying high-efficiency CSP systems. The water vapor with condensed
droplets that emanate from power station cooling systems represents the potential for low-
temperature waste heat to be used as an addition heat source that can be converted to useful work
or power. Exhaust heat is represented by 𝑄̇𝑜𝑢𝑡 flowing from of the lower side of the T-s diagram
Fig. 6.4, where processes 1-2 and 3-4, represented by vertical lines, closely resemble that of the
Carnot cycle. For this cycle, vapor is prevented from the superheated region after the expansion in
the turbine, which reduces the energy removed by the condensers. The actual vapor power cycle
differs from an ideal Rankine cycle due to irreversibilities inherent within the components by fluid
friction and heat losses to the ambient. Additionally, fluid friction also facilitates pressure drops
(or head losses) within the heat input stage (generally also known as the “boiler”), the condenser,
and system piping, which can result in steam leaving the boiler at a lower pressure, where heat
losses reduce work output. For the Rankine cycle, Fig. 6.4 presents these four principal stages:

 Process 1-2: Compression of the liquid working fluid to high pressure. The working fluid is
pumped from low to high pressure. As the fluid remains liquid at this stage, the pump requires
relatively low input energy.

 Process 2-3: Heating and subsequent vaporization of the working fluid that is driven by a heat
source. The high pressure liquid enters a heat input stage where it is heated at constant pressure
by an external heat source to become a dry saturated vapor. The input energy required can be
calculated by an h-s chart (Mollier diagram), or numerically using pre-defined tables.

 Process 3-4: Expansion through a turbine to a lower pressure for generating mechanical work.
The dry saturated vapor expands through a turbine, generating power, which decreases the
temperature and pressure of the vapor.

114
 Process 4-1: Cooling of the working fluid back to its initial state. The wet vapor then enters a
condenser where it is condensed at constant pressure to become a saturated liquid.

Figure 6.4. Ideal Rankine cycle components and T-s diagram [315].

In an ideal Rankine cycle the pump and turbine would be isentropic (i.e., the pump and turbine
would not generate entropy and therefore maximize the net work output). The efficiency of the
Rankine cycle is limited by the high heat of vaporization of the working fluid. Unless pressure and
temperature reach super critical levels within the heat input stage, the temperature range the cycle
can operate over is quite small. For water as the working fluid, steam turbine entry temperatures
are typically around 565 °C and steam condenser temperatures are around 30 °C [315]. This gives
a theoretical maximum Carnot efficiency for the steam turbine alone of about 63% compared with
an actual overall thermal efficiency of up to 42% for a modern coal-fired power station. This low
steam turbine entry temperature (compared to a gas turbine system) is why the Rankine cycle is
often used as a bottoming cycle to recover otherwise rejected heat in combined-cycle gas turbine
power stations [203]. Additionally, subcritical steam Rankine cycles used in solar facilities such
as Solar One or SEGS [316] operate within the efficiency range of 37-42% [66, 317]. This type of
cycle, used for CSP parabolic troughs, have typically been limited to turbine inlet temperatures of
less than 400 ᵒC due to limitations of oil, and relatively low solar flux levels provided by
concentrating optics [203]. The utility of molten salt used as a working fluid systems that do not
use an intermediate HTF (such as those that generate steam directly) have previously been explored
[318]. These systems have been shown to raise turbine inlet temperature above 400 ᵒC for trough
systems [318, 319].
Cooling towers operate as large heat exchangers by absorbing the latent heat of
vaporization of the working fluid, while simultaneously evaporating cooling water to the
atmosphere. In traditional Rankine cycles, water is typically used due to its favorable properties,
such as its non-toxic and unreactive chemistry, abundance, low cost, and favorable thermodynamic
properties. By condensing steam vapor to a liquid, the pressure at the turbine outlet is lowered and
the energy required by the feed pump consumes only 1% to 3% of the turbine output power where
these factors contribute to a higher overall efficiency [203]. Gas turbines traditionally experience
entry temperatures approaching 1500 °C. However, thermal efficiencies of actual large steam
power stations and large modern gas turbine stations have been found to be similar [203].
For CSP, molten-salt power tower plants are limited to receiver outlet temperatures of
approximately 565–600 ᵒC and employ heat exchangers to produce superheated steam at
approximately 540 °C and 130 bar, with reheated steam at 538 ᵒC and 28 bar [202]. The result in

115
gross thermal efficiencies are typically 42% with wet cooling [202]. Typical steam Rankine cycles
are limited to operational temperatures of up to approximately 600 °C, where higher efficiency
can be achieved with supercritical steam Rankine cycles, with temperatures up to 760 °C [139].
Supercritical steam cycles have the potential to operate at higher temperatures and pressures,
increasing the thermal efficiency to above 45% [250, 320].
Operating with steam in the supercritical regime allows for higher temperatures and fluid
behavior that shares the compressibility traits of a liquid (reducing compressor work and enhancing
regenerative heat exchange) while avoiding complications due to multiphase heat exchangers.
However, concern has been expressed when steam temperatures exceed 627 ᵒC due to the
limitations of standard ferritic steels; these higher temperatures are expected to require the use of
high-nickel composition alloys [203, 320, 323]. Organic Rankine Cycles (ORC) have been studied
in-depth for their use in medium-temperature, and waste heat recovery applications [321, 322],
though they typically have efficiencies of approximately 10–20%, and are most often used as
bottoming cycles when the exit stream from a high temperature cycle still possesses high
availability.
Additionally, alkali metals have been proposed for high-temperature topping cycles in a
Rankine–Rankine combined cycle configuration [220]. This system provides good temperature-
matching with a bottoming steam cycle with high heat transfer coefficients due to liquid metal
phase change in the condenser. Thermal efficiencies of an alkali metal Rankine topping cycle and
steam Rankine bottoming cycle have been shown to reach 56% for 1000 ᵒC maximum working
fluid temperature, with the potential to reach 60% [203]. However, turbomachinery for metal vapor
cycles is estimated to be large, with a 50 MW turbine tip diameter of about 3.9 m for potassium
and 2.8 m for rubidium [220]. Finally, although organic Rankine cycles have been well studied,
primarily for medium temperature systems [321], they typically only have seen efficiencies of
approximately 10-20%, where they are most often used in bottoming cycles when the exit steam
from a high temperature cycle still possesses high availability [320].

6.4 Brayton Cycle

6.4.1 Overview & Theory

A Brayton cycle behaves similar to a Rankine cycle with the exception that the working fluid
remains in the gas phase throughout the cycle [203]. Brayton cycles are able to operate at much
higher temperatures, which can increase potential thermal efficiencies. Innovative receiver
concepts have been proposed and demonstrated which can heat gases to temperatures above 1000
ᵒC [68, 324]. However, the compression stage of gas-Brayton cycles requires much more power
[203]. Although CSP trough systems typically operate at approximately 400 ᵒC and Rankine cycle
power towers operate at temperatures up to approximately 540 ᵒC, Brayton thermodynamic cycles
can operate at temperatures approaching 1,000 ᵒC, offering development opportunities for greatly
increased efficiency [66]. Higher efficiencies on the order of 50% are predicted for advanced, high-
power, multiple-reheat, helium Brayton cycles which could operate with a turbine inlet
temperature from 750 to 850 ᵒC [203]. These temperatures are significantly higher than those used
in subcritical steam Rankine cycles, but are achievable by current high-concentration solar tower
technology. Helium possesses good heat transfer characteristics which also could be beneficial in

116
the design of the solar receiver. Angelino and Invernizzi [220] note that the closed helium cycle
could be useful in cogeneration applications, as the rejected heat is typically available at relatively
high temperatures. In 2002, a hybrid open solar Brayton cycle was operated for the first time
consistently and effectively with relevant papers published, in the frame of the EU SOLGATE
program [349]. The air was heated from 570 K to over 1000 K into the combustor chamber. Further
hybridization was achieved during the EU Solhyco project running a hybridized Brayton cycle
with solar energy and Biodiesel only [350]. This technology was scaled up to 4.6 MW within the
project Solugas located near Seville where it is currently demonstrated at pre-commercial scale.
The premise of a Brayton cycle system is to extract energy from a flowing gas and fuel to
generate useable work to power jet-powered aircraft and utility power block generators by
providing thrust to a turbine. The cycle as shown in Fig. 6.5a can employ a heat exchanger or
regenerator for heat removal, or it can be absent, as shown in Fig. 6.5b where the hot gas is
exhausted directly to the ambient.

a. b.

Figure 6.5. a. Closed loop ideal Brayton cycle and b. illustration of an open Brayton cycle system
[326].

The most basic steps within a Brayton cycle in extracting energy is compression of flowing air,
combustion, and then expansion of that air to create work and also power the compression at the
same time. Although many irreversibilities exist within real systems, an ideal approach can be
employed, Fig. 6.6 to quickly assess power production and efficiency, based on the following state
processes:

 Process 1-2: Inlet gas enters and contacts the compressor which causes the pressure and
temperature to dramatically rise. The rise in pressure comes from work being facilitated by the
compressor which packs the gas into the mixer/combustion chamber. The rise in pressure
accelerates a rise in temperature of the gas molecules due to the volume staying constant. For
an ideal process (PV = nRT), entropy is assumed to be relatively constant (isentropic process)
due to the flow of the gas.

 Processes 2-3: As the gas is compressed into the combustion chamber, it is typically mixed
with a fuel to facilitate ignition, however for CSP applications, this process can be partially or
fully omitted as hot gas is allowed to enter from a high-temperature receiver. For combustion

117
 processes, once the gas-fuel mixture is ignited a steep rise in temperature and entropy is
facilitated. The energy from the chemical bonds in the fuel are broken due to ignition, which
is highly exothermic which raises entropy and raises the temperature due to increased ambient
energy from the reaction.

 Process 3-4: The pressurized (and elevated temperature) fuel and air leave the combustion
chamber to an expansion chamber, where a rapid drop in pressure is facilitated due to an
increased volume change. The energy from combustion (increased P and T) is used to exert a
force on a turbine/generator system and to the compressor for continuous Brayton cycle
operation. The remaining energy from this process can be leveraged as a higher energy level
input in an intermediate process for enhancing efficiency, or will be lost to surroundings
(isobaric process).

 Process 4-1: Cooler exhaust gas from the turbine may enter into a heat exchanger which
facilitates heating of the working fluid from a high-temperature receiver or from TES.
Additionally, this process may be omitted if the working fluid is directly heated from a
receiver.

a. b.
Figure 6.6. Ideal Brayton Cycle a. P-V Diagram and b. T-S Diagram.

The thermal efficiency of the cycle can be determined from the net output work and the heat
absorbed. Tracing the path shown around the cycle from a-b-c-d, the first law, in terms of unit
mass, provides:
∆𝑢𝑎−𝑏−𝑐−𝑑−𝑎 = 0 = 𝑞2 + 𝑞1 − 𝑤 (6.17)

Here ∆u is approximated as zero where u is a function of state, and where any cycle returns the
system to its starting state [326]. The net work is therefore:

𝑤 = 𝑞2 + 𝑞1 (6.18)

where q1 and q2 are defined as heat received and rejected by the system respectively. For a constant
pressure, quasi-static process the heat exchange per unit mass is defined by writing the first law in
terms of enthalpy:
𝑑ℎ = 𝐶𝑝 𝑑𝑇 = 𝑑𝑞 (6.19)

118
In general, heat exchange can be expressed in terms of enthalpy differences between respective
states, where the working fluid can be idealized as a perfect gas with constant heat capacity values.
Therefore, the heat addition from the combustor can be expressed as:

𝑞2 = ℎ𝑐 − ℎ𝑏 = 𝐶𝑝 (𝑇𝑐 − 𝑇𝑏 ) (6.20)

Similarly, the heat rejected can be expressed as:

𝑞1 = ℎ𝑎 − ℎ𝑑 = 𝐶𝑝 (𝑇𝑎 − 𝑇𝑑 ) (6.21)

The net work per unit mass is provided by:

𝑁𝑒𝑡 𝑊𝑜𝑟𝑘 𝑝𝑒𝑟 𝑈𝑛𝑖𝑡 𝑀𝑎𝑠𝑠 = 𝑞1 + 𝑞2 = 𝐶𝑝 [(𝑇𝑐 − 𝑇𝑏 ) + (𝑇𝑎 − 𝑇𝑑 )] (6.22)

The Brayton cycle thermal efficiency can then be expressed in terms of state temperatures:
𝑇
𝑁𝑒𝑡 𝑊𝑜𝑟𝑘 𝐶𝑝 [(𝑇𝑐 −𝑇𝑏 )−(𝑇𝑑 −𝑇𝑎 )] (𝑇𝑑 −𝑇𝑎 ) 𝑇𝑎 ( 𝑑 −1)
𝑇𝑎
𝜂= = = 1 − (𝑇 −𝑇 = 𝑇
(6.23)
𝐻𝑒𝑎𝑡 𝐼𝑛 𝐶𝑝 [𝑇𝑐 −𝑇𝑏 ] 𝑐 𝑏) 𝑇𝑏 ( 𝑐 −1)
𝑇𝑏

To assess temperature relationships between the states within the cycle it is important to note that
points a and d, as well as states b and c, operate on a constant pressure processes, where Pa=Pd,
Pb=Pc. The other processes of the cycle are adiabatic and reversible, where:

𝑃𝑑 𝑃 𝑇 𝛾⁄(𝛾−1) 𝑇 𝛾⁄(𝛾−1)
= 𝑃𝑎 → ( 𝑇𝑑 ) = (𝑇𝑎 ) (6.24)
𝑃𝑐 𝑏 𝑐 𝑏

Accordingly, Td/Tc=Ta/Tb and Td/Ta=Tc/Tb. Using this relation in the expression for thermal
efficiency, the above equation yields an expression for the thermal efficiency of a Brayton cycle:
𝑇 𝑇𝑎𝑡𝑚
𝐼𝑑𝑒𝑎𝑙 𝐵𝑟𝑎𝑦𝑡𝑜𝑛 𝐶𝑦𝑐𝑙𝑒 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦: 𝜂𝐵𝑟𝑎𝑦𝑡𝑜𝑛 = 1 − 𝑇𝑎 = 1 − 𝑇 (6.25)
𝑏 𝑐𝑜𝑚𝑝,𝑜

The efficiency of the Brayton cycle be determined based of the temperature ratio across the
compressor, TR=Tb/Ta, the pressure ratio PR, and the relation for an adiabatic reversible process:
1 1
𝜂𝐵𝑟𝑎𝑦𝑡𝑜𝑛 = 1 − 𝑇𝑅 = 1 − 𝑃𝑅(𝛾−1)⁄𝛾 (6.26)

As prescribed by Eqn. 6.26, to achieve a high thermal efficiency, with system parameters
prescribed by [326], the pressure ratio of the cycle should be increased as demonstrated by Fig.
6.7.

119
 Figure 6.7. Ideal Brayton Cycle Compressor Pressure Ratio versus Thermal Efficiency [326].

6.4.2 Supercritical Brayton Cycles

Supercritical operation can permit an enhanced capture and utilization of heat, taking advantage
of both latent and sensible HTFs in the two-phase regime as well as in the supercritical regime,
while also reducing the required volume by taking advantage of high compressibility.
Additionally, for supercritical systems storage performance and pressures can be optimized by
judicious selection of fluids, such as sCO2, with the following properties:

1. High latent heat of vaporization

2. High specific heat
3. High critical point
4. Low vapor pressure

For supercritical CO2 systems, a heat source or CSP receiver can be incorporated within the loop
as illustrated by the simple sCO2 layout and T-s diagram provided by Goswami [332], Fig. 6.8.

a. b.
Figure 6.8. Simple sCO2 Brayton cycle a. component layout and b. T-s diagram [332].

Supercritical gases can be advantageous working fluids for use in power generating turbines as
they offer high efficiency with a compact footprint that can be matched to many different heat
sources in CSP. In addition to increasing interest in nuclear and renewable energy, sCO2 is
becoming an important commercial and industrial solvent due to its advantages in chemical

120
extraction, low toxicity, stability and environmental impact. Previous research has shown promise
for sCO2 transport in pipelines, in addition to energy storage, as illustrated in Fig. 6.9 [327].

Figure 6.9. sCO2 phase diagram with various CCS operations with superimposed
pressure/temperature operating domains [327].

In an sCO2 system, the fluid is compressed where the high turbine inlet temperature can be utilized
with less material issues compared with steam Rankine cycles. However, the cycle pressure ratio
of the sCO2 Brayton cycle is much smaller than the steam Rankine cycle where the turbine outlet
temperature is relatively high. Therefore, a large amount of heat must be recuperated to increase
thermal efficiency, where the recuperation process greatly influences the thermal efficiency.
The CO2 critical condition is achieved at 30.98 ᵒC and 7.38 MPa where the fluid becomes
more incompressible near the critical point as illustrated by Fig. 6.10 based on Eqn. 4.7.

Figure 6.10. CO2 compressibility factor as a function of temperature [328].

In the supercritical region, fluids exhibit compressibility and heat transfer characteristics more like
liquids while still reaching high temperatures. This allows for reduced compressor work and more

121
efficient regenerative heat exchange [328]. CO2 behaves like an ideal gas when the compressibility
factor (molecular volumetric ratio of a fluid compared with ideal gas) is unity and is considered to
be an incompressible fluid when it is zero. For operation near the critical point, the compressibility
factor decreases to values between 0.2-0.5 where the compression work can therefore be
substantially decreased [328]. Additionally, previous research within the aerospace and nuclear
industries has demonstrated unique thermodynamic performance enhancements of Brayton
systems with the utility of sCO2 as the working fluid. As shown in Fig 6.11, for constant pressure
conditions, heat capacity and enthalpy of CO2 has remarkable thermodynamic property advantages
over other readily available gases for employment in high temperature power systems [329].

a. b.
Figure. 6.11. Comparisons of gases under constant pressure for a. specific heat and b. enthalpy
[329].

As illustrated by Fig. 6.12, during operation in a CSP system, an sCO2 bottoming cycle is first
pumped up to a supercritical pressure, and then heated by a gas or heat from a solar receiver. The
heated sCO2 gas will then expand in the turbine producing work, where the vapor is then
discharged from the turbine outlet to be cooled and condensed within a condenser (typically a gas
cooler). The internal heat exchanger (regenerator) is typically included to optimize the
performance. Provided that the compressor inlet conditions are kept in the vicinity of the critical
point the specific compressor work is low and the compressor absorbs a small part of the work
produced by the turbine [330].

122
 a. b.
Figure 6.12. a. Carbon dioxide power system schematic and b. Carbon dioxide Brayton cycle T-s
diagram [331].

Additionally, a notable thermodynamic characteristic of sCO2 is its specific heat with respect to
temperature as illustrated by Fig. 6.13. During operation the cold-side recuperator can have
specific heat values that can be as high as two to three times that of the hot-side recuperator, which
can have a negative impact on performance [328]. Therefore, design measures can be taken to split
the sCO2 flow to compensate for this specific heat difference to maximize recuperation for a
recompression design, which reduces waste heat and improves thermal efficiency [328].

a. b.

Figure 6.13. Specific heat and density properties of sCO2 around the critical point [332].

The region near the CO2 critical point is marked by rapid fluctuations in thermodynamic properties
which may present a challenge in the design and operation of turbomachinery, particularly for a
transient heat source from concentrated solar energy [203]. An incentive for using sCO2 in turbine
power cycles is also based on its favorable thermal stability compared to steam which allows for
much higher power output in a much smaller package than comparable steam cycles [64]. The
sCO2 cycle is inherently efficient due to its operation near its respective critical point and the

123
ability of the working fluid to tolerate very high inlet temperatures. The high fluid density of sCO2
enables extremely compact turbomachinery designs that can be highly efficient with simple, single
casing body designs. Fig. 6.14 demonstrates the relative size difference for two 10 MWe systems,
where the overall size of sCO2 turbo-machinery is significantly smaller than either conventional
steam or helium turbines [338]. Additionally, due to its relative smaller size, many of the
components for the sCO2 system can be manufactured with advanced adaptive manufacturing
technology [339].

Figure 6.14. Comparison of Echogen 10 MWe sCO2 turbine with 10 MWe traditional steam turbine
[338].

More conventional steam turbines require multiple stages (e.g. low, medium and high-pressure)
with associated casings, which can lead to a corresponding increase in inlet/outlet systems
packaging complexity and O&M issues [338]. sCO2 is effective at capturing waste heat from
sources that have an approximately constant heat capacity, such as turbine exhaust or other hot
gases due to the character of its heat capacity in the supercritical region. This can provide superior
matching to the heat source temperature profile compared with boiling processes with other
working fluids (steam or organic working fluids) used in organic Rankine cycle systems [338].

Figure 6.15. a. Steam and b. supercritical CO2 working HTFs during heating in an exhaust heat
exchanger [338].

As shown in Fig. 6.15 in comparing gas heat exchanger performance between CO2 and steam that
undergoes constant-temperature phase change, Persichilli et al. demonstrated the pinch point

124
phenomenon, which limits the maximum fluid temperature and resulting cycle efficiency [338].
The investigators determined that for steam-based systems, the boiling process limits the
maximum fluid temperature and can require multiple heat recovery steam generators (HRSG) to
get close to the same CO2 gas temperature thereby reducing heat exchanger effectiveness.
Angelino found that single-heating sCO2 cycles above 650 ᵒC have a higher efficiency than
reheat steam Rankine cycles [220]. Dostal et al. showed a thermal efficiency of 46.07% for an
sCO2 Brayton cycle with a turbine inlet temperature of 550 ᵒC, and 49.25% for a helium Brayton
cycle with turbine inlet temperature of 880 ᵒC [334]. The volumetric flow rate for sCO2 is much
less than that for helium by approximately a factor of 5 [203], leading to relatively smaller
turbomachinery requirements. Moisseytsev and Sienicki examined an sCO2 cycle for a lower-
temperature nuclear-driven process which operated between 31.25 ᵒC/7.4 MPa and 470 ᵒC/20.0
MPa with a thermal efficiency of 39.1% [334]. An increase in thermal efficiency to 43.1% was
found by reducing the minimum temperature and pressure to 20 ᵒC/5.75 MPa so the cycle operated
as a condensation cycle. However, the low condensation temperature requires heat rejection to a
relatively low temperature that may not be possible in environments where the ambient
temperature is high. The sCO2 Brayton cycle has been considered for use in CSP in studies of
transient behavior by a cyclic, solar-driven heat source [335]. Garg et al. [336] concluded that by
operating in the supercritical regime, the same thermal efficiency can be achieved with a lower
maximum temperature, suggesting the potential for cost savings. Ma and Turchi suggested a
modular tower design to capitalize on small sCO2 turbomachinery where the entire power block
would be placed on a power tower rather than transport high-pressure fluid to an externally located
power block or incur large financial costs with a single large tower [261]. Additionally, Singh et
al. [337] have also found promise for employment of sCO2 in parabolic trough concentrators.
Challenges for using sCO2 for CSP applications are generally focused on materials that can
handle elevated temperatures/pressures, manufacturing turbo machinery, valves, pumps, seals, and
high costs. Compatible heat exchangers that can effectively transfer heat from a receiver HTF to a
sCO2 working fluid is also a significant challenge that has garnered much research over the last
decade [340]. An analysis by Dunham and Iverson [203] performed a review of several high-
efficiency thermodynamic cycles and the applicability to CSP systems, where they found in
comparison to He-Brayton, regenerated CO2 Brayton, CO2 recompression Brayton and CO2-ORC
combined cycles that steam Rankine cycles showed higher thermal efficiencies at temperatures up
to 600 ᵒC. Above this temperature, CO2 recompression Brayton cycles were shown to have a very
high superior thermal efficiency, potentially even exceeding 60% at 30 MPa maximum pressure
and above 1000 ᵒC maximum temperature with wet cooling [203]. For these cycles, HTF salts
have also been investigated, such as the work by Forsberg et al. [341], who proposed high
operating temperature molten salts (>700 ᵒC) for use in power towers with direct thermocline
TESS and graphite as a filer material. The multi-reheat Brayton cycle utilized helium and nitrogen
as the working fluid and was able to achieve operating temperatures between 700 ᵒC and 1000 ᵒC
[341]. However, liquid metals have the capability of operating at low pressures and potential for
reaching the temperatures required for next-generation Rankine or supercritical Brayton power
cycles.

125
6.4.3 Brayton Cycle Components & Advanced Designs

The output power of a Brayton Cycle can be significantly improved through the addition of heat
transfer components that can facilitate increased temperatures, pressures and subsequent thermal
efficiency. Several sCO2 Brayton cycle layouts have been previously investigated [220, 225] to
improve efficiency with respect to particular included processes:

1. Reheat
2. Recuperation and Recompression
3. Intercooling and Regeneration

6.4.3.1 Reheat

To bring the average temperature closer to the peak temperature, heat can be added by a reheat
stage. As shown in Fig. 6.16, a reheater is a heat exchanger that increases the power output without
increasing the maximum operating temperature. Here, excess working fluid is extracted and fed
into a second stage combustor and turbine where the turbine outlet temperature is increased with
reheat (T6 > T4). This increases the potential for regeneration which can in turn significantly
increase thermal efficiency [346].

a. b.

Figure 6.16. Ideal Brayton Cycle Reheat Process with heat exchanger (HEX) with a. Component
Diagram and b. T-s Diagram [246].

In this layout the working fluid expands through a series of turbines, which then passed through a
second combustion chamber before expanding to ambient pressure through a final set of turbines.
This has the advantage of increasing power output for a given compression ratio without exceeding
any metallurgical constraints (at approximately 1000 °C [345]). Reheat is most often used to
improve the specific power (per throughput of air), and is usually associated with a drop in
efficiency [345]. However, this process does not directly increase the efficiency of the cycle.

6.4.3.2 Recuperation & Recompression

The efficiency of a Brayton power cycle can also be improved by means of increasing the pressure
ratio and utility of a recuperator. Recuperator performance can be expressed in terms of

126
effectiveness (percentage of possible heat transfer achieved) or pinch (minimum temperature
differences at the two ends of the recuperator) [347]. Within a recompression cycle that employs
a recuperator, this configuration differs from the simple Brayton cycle in that before cooling, some
fraction of the flow does not pass through the precooler and the main compressor. Instead it is
recompressed and then rejoins the main flow between the two parts of the divided recuperator, Fig.
6.17a. It is assumed that at this point the two flows are at the same temperature. The flow, which
passes the low temperature recuperator should have the same pressure and temperature as the other,
where both streams are later mixed again into one stream. This prevents the pinch point issue due
to lower mass flow at the high pressure side of the low temperature recuperator since the heat
capacity mass flow, weighted on both sides, are equal. This system will therefore reject less heat,
and because re-compressor input work is lower than saved heat, thermal efficiency is improved
[348].
Increasing the pressure ratio also has the capability of increasing the compressor discharge
temperature, which can cause the temperature of the gasses leaving the combustor to exceed the
metallurgical limits of the turbine. Additionally, the diameter of the compressor blades becomes
progressively smaller in higher pressure stages of the compressor. This causes the gap between the
blades and the engine casing to increase in size as a percentage of the compressor blade height as
the blades get smaller in diameter. Therefore, a greater percentage of the compressed air can leak
back past the blades in higher pressure stages, which can cause a drop in compressor efficiency,
and is most likely to occur in smaller gas turbines (since their blades are inherently smaller to begin
with).
In an analysis by Ahn et al. [328] who investigated thermodynamic properties of sCO2
around the critical point, the authors’ computational results suggest that for increasing pressures,
the density was found to increase substantially (Fig. 6.17) which can influence the minimum
pressure for turbine operation as compared to a traditional Gas-Brayton cycle (approximately
7,400 kPa compared to approximately 100 kPa) [347].

a. b.
Figure. 6.17. sCO2 recompressing cycle with low and high-temperature recuperators [328].

127
If the Brayton power system is run at a low enough pressure ratio, the turbine outlet gas after the
second turbine may still be hotter than the compressed inlet gas after the first turbine. Here, a heat
exchanger can be used to transfer thermal energy from the turbine exhaust to the already
compressed gas before it enters the combustion chamber. The thermal energy transferred is
effectively re-used, thus increasing efficiency. However, this form of heat recycling is only
possible if the engine is run in a low efficiency mode with a low pressure ratio. Note, that
transferring heat from the 2nd turbine outlet to the inlet of the first turbine would reduce efficiency,
as a hotter inlet gas would require more volume and in turn more work for the compressor turbine
[328].

6.4.3.3 Intercooling & Regeneration

A Brayton Cycle with Intercooling uses two or more compression stages with one or more
intercoolers, as shown in Fig. 6.18. With this configuration, although the power requirement for
compression is reduced, the heat quality also increases. For this layout, a regenerator is used as a
heat exchanger, used to recover heat from the turbine exhaust that would otherwise be wasted.
This is accomplished by connecting the relatively hot turbine effluent gas with a relatively cool
compressor outlet gas.

a. b.
Figure 6.18. Ideal Brayton Cycle with Intercooling and heat exchanger (HEX) with a.
component diagram and b. T-S Diagram [246].

Additionally, sCO2 is often more sensitive to the critical point of working fluids within Brayton
cycle systems than other working fluids [420] since compressor work is significantly decreased
slightly above the critical point due to high density CO2 fluid near the boundary between the
supercritical state and its subcritical state. To address this challenge and increase the efficiency of
sCO2 Brayton systems, a second compressor (or recompressor) can be added to the simple Brayton
cycle [421]. Although the two parallel compressors have different inlet conditions, they must
operate at nearly the same pressure ratio to avoid impacting the performance of the other
compressor, where a relatively small difference in pressure ratio can overload one of the
compressors and shutdown the system. The addition of a regenerator and recompression within a
CSP Brayton system has previously been explored by Dunham and Iverson [203] where the
investigators first assessed a closed regenerative Brayton cycle with one stage of intercooling, Fig
6.19a. Here, a fluid at the minimum system temperature enters the first-stage compressor at state

128
(1) and is compressed to a higher pressure and temperature, state (2). A heat exchanger (HEX)
removes heat from the fluid before it enters the second-stage compressor at state (3).

a.

b.
Figure 6.19. Brayton thermodynamic schematics with a. regeneration and b. recompression [203].

This intercooling process serves two primary purposes: it reduces the total work required for
compression as well as reducing the compressed fluid outlet temperature which allows for more
substantial regeneration, Fig. 6.19b. The compressed gas then enters the cold side of the
regenerator at state (4) where it is preheated before entering the receiver at state (5). Now at its
highest temperature, the compressed gas enters the turbine at state (6) where it is expanded to low
pressure. At state (7), the hot turbine outlet stream enters the hot side of the regenerator, preheating
the cold-side fluid. Finally, the gas enters a heat exchanger at state (8) through which it is returned
to the minimum cycle temperature. In their investigation Dunham and Iverson [203] suggested the
following compression ratio of each compressor in a two-stage intercooled process with known
minimum and maximum pressures and a constant pressure drop within the intercooler, Eqn. 6.27.

2
𝑃𝑙𝑜𝑠𝑠,𝐻𝐸𝑋 +√(𝑃𝑙𝑜𝑠𝑠,𝐻𝐸𝑋 ) −4𝑃𝑚𝑖𝑛 𝑃𝑚𝑎𝑥
𝑃𝑟,2−𝑠𝑡𝑎𝑔𝑒 = (6.27)
2𝑃𝑚𝑖𝑛

Additionally, from Fig. 6.18a the investigators also suggest using Eqn. 6.28 for regenerator
effectiveness, calculated using the state enthalpies [333].

ℎ5 −ℎ4
𝜀𝑟𝑒𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑜𝑟 = ℎ (6.28)
7 −ℎ(𝑇4 ,𝑃7 )

129
where the term ℎ(𝑇4 ,𝑃7 ) is necessary for supercritical applications, though can commonly be
approximated to be the same as ℎ4 under ideal gas systems [422].
As illustrated in Fig. 6.18b the recompression Brayton cycle is typically designed to have
a fraction (~10-40%) of hot steam returned to the minimum cycle temperature while the remaining
flow bypasses the cooler and is recompressed from the higher temperature [203]. Previous research
of the recompression cycle has primarily been focused on sCO2 [264], especially as the operation
of the compressor near the critical point reduces the work required for gas compression. This is
due to the fluid’s unique thermodynamic properties as it transitions to a supercritical state [203].
Since CO2 behaves less like an ideal gas than He under most conditions, the efficiency is more
significantly-impacted by pressure. In their investigation, Dunham and Iverson [203] also analyzed
a regeneration Brayton cycle system under dry-cooling and wet cooling conditions. Their results
suggested that a high-temperature CO2- Brayton cycle shows more resilience to the effects of dry-
cooling than does the He-Brayton cycle, although the thermal efficiencies were found to be lower.
However, the authors suggested that more careful consideration of pressure may be required during
the design of a solar receiver for CO2 systems than with He systems. Performance results for a
CO2 recompression Brayton cycle under wet-cooling conditions revealed that the recompression
cycle showed the potential for very high thermal efficiencies, particularly at high temperatures and
pressures. For 20 MPa conditions, the compressor inlet condition was supercritical for all
temperatures with wet-cooling and was supercritical at a temperature of 1000 ᵒC for dry-cooling.
At 30 MPa, a 50% thermal efficiency was reached with a maximum temperature slightly above
650 ᵒC for wet-cooling, and slightly under 800 ᵒC for dry-cooling [203]. Their results also found
for the regenerated He-Brayton cycle, under wet-cooling conditions, that as long as the minimum
pressure is free to vary, the maximum pressure has relatively minimal impact on the thermal
efficiency. This was found under operating conditions where the thermal efficacy rose from 55.0%
at a maximum pressure of 10 MPa to only 55.7% at 30 MPa for a maximum temperature of 1100
ᵒC. The maximum temperature has a significant impact on the system, causing the cycle thermal
efficiency to rise from 29.3% at 500 ᵒC, to 55.7% at 1100 ᵒC and 30 MPa maximum pressure. This
suggests that lower system pressures may be employed with relatively minimal effects on
performance, providing fewer constraints on thermal optimization of the solar receiver [203].

6.4.3.4 Combined Cycle

Another method for enhancing sCO2 power cycle performance is through the utility of a combined
cycle which utilizes multiple thermodynamic cycles: one primary high-temperature cycle
(topping) and one or more lower-temperature cycles (bottoming) driven by the heat rejected from
the higher-temperature cycle. While significant steps have been made towards high-efficiency
single cycles, it has been widely believed that operation above 50% thermal efficiency [68, 308]
may require the implementation of binary or combined cycles [313], though it has been suggested
that advanced sCO2 and helium Brayton cycles could potentially reach to levels greater than 50%.
Furthermore, combined cycles have been shown to be techno-economically effective, despite
increased capital costs [310]. Brayton-topped, combined cycles allow for optimum use of highly
concentrated solar energy where the ideal operating temperature is beyond what can be used in
subcritical Rankine cycles [314]. An sCO2 combined cycle that’s been previously studied is a
configuration that consists of a topping Brayton cycle with a bottoming ORC for waste heat

130
recovery [342, 343]. A study with an air-ORC/sCO2 combined cycle showed that 50% thermal
efficiency was possible for gas turbine inlet temperatures above 1027 ᵒC, with the best
combination at this temperature being air-toluene with an efficiency of 53.75% [342]. The peak
temperature of toluene in the bottoming cycle was found to be 297 ᵒC at the turbine inlet. A high-
temperature combined cycle based on a closed-loop CO2 Brayton cycle and driven by concentrated
solar radiation was examined and found to produce a thermal efficiency of 47.51% for a turbine
inlet temperature of 827 ᵒC when combined with an isopentane bottoming cycle [343]. A variety
of working fluids proposed for use in low-temperature or bottoming cycles were compared by
Chen et al. [321]. It was determined that the critical temperature and classification as a “dry” or
“wet” fluid (defined by the slope of the T–s diagram) were significant in determining the fluid's
suitability for a given application. Further, dry fluids actually perform better when not superheated
at the turbine inlet, which can reduce the need for additional heated sections. Johnson examined
power cycle performance with a nuclear source and concluded that a single sCO2 or cascaded sCO2
cycle had benefits over a combined helium Brayton and supercritical steam Rankine cycle in terms
of performance and turbo-machinery size [344]. For a topping cycle turbine inlet temperature of
850 ᵒC, Johnson calculated the net thermal efficiency to be 49.8% for a cascaded sCO2
configuration [203].

6.4.3.5 Overspray

For traditional Brayton cycle power systems, an overspray method can also be utilized which
occurs during a first stage compressor, where a working fluid is injected into the compressor, thus
increasing the mass-flow rate which increases the turbine output power significantly, thereby
reducing compressor outlet temperatures [3]. For a second stage compressor with non-gas working
fluids, the fluid is completely converted to a vapor form offering some intercooling via its latent
heat of vaporization.

6.5 Stirling Cycles

Stirling cycles are used frequently during CSP dish applications [352]. In a Stirling cycle, air,
helium or hydrogen are used as working fluids where all processes of this cycle are assumed to be
reversible, therefore when the working fluid is heated the engine produces work or power.
Conversely as illustrated by Fig. 6.20, if the engine is driven by a motor, and the process curves
in the T-s and P-v diagrams run counter-clockwise, the Stirling cycle operates as a refrigerator
and/or heat pump. For the Stirling cycle, Carnot cycle’s compression and expansion isentropic
processes are replaced by two constant-volume regeneration processes. During the regeneration
process heat is transferred to a thermal storage device (regenerator) and is transferred back to the
working fluid in another part of the cycle. Here, the regenerator can be a wire, ceramic mesh or
any kind of porous plug with a high thermal mass [352].
Unlike other combustion engines, a Stirling cycle does not exchange the working gas in
each cycle, therefore the gas is permanent. The heat is supplied externally, and can have the
potential of achieving higher thermal efficiencies than even Otto and Diesel engines, since heat
transfer occurs at constant temperatures. This type of heat transfer can be particularly suitable for
systems with latent HTFs since they transfer thermal energy under constant temperature operation.

131
In this cycle, the pressure changes are very smooth and its torque is uniform where the system does
not require any valves. Therefore, Stirling cycle systems are quiet and require little maintenance.
To achieve competitive efficiency, the system needs to work at high pressures which could
contribute to sealing issues. The temperature difference TL - TH should be maintained high for
acceptable thermal efficiencies, this results in large thermal stresses in the cylinder (hot and cold
ends). Consequently, high strength materials should be used such as Inconels and Haynes 230. The
working fluid has to be an inert gas, where Helium could be ideal for this cycle due to its high heat
capacity and low molecular mass, which lead to enhanced heat transfer [353].

Figure 6.20. Stirling cycle operation and P-V, T-s thermodynamics [344].

6.6 Other CSP Cycles

Novel thermodynamic cycles not traditionally used in either CSP or fossil fuel plants, have also
been theoretically explored for CSP application, though primarily for use in combined cycles
[203]. A fringe cycle has been explored using a water-ammonia HTF where this cycle utilizes
differing boiling points to fit fluid thermodynamics to the heat source, which allows for an increase
in overall efficiency as high as 52% [355]. This type of cycle however generally has lower
operating temperatures and has been considered for bottoming cycles, with working fluid
properties tuned to the topping cycle heat rejection [356]. However, further study by Chen et al.

132
[321] found compared to an ORC cycle, this type of bottoming cycle requires more complex
equipment.
Kribus proposed a system involving a triple “cycle” for solar energy conversion that is
much like a typical combined cycle but with an additional energy conversion above the topping
cycle [203]. Feeding the topping cycle is the waste heat from an MHD cycle, which uses extremely
hot ionized air, heated to approximately 2000 ᵒC to directly induce an electric current using
magnets [357]. Very high efficiencies are possible in this configuration, theoretically approaching
70%. However, concentration ratios of approximately10,000 suns are also needed, requiring a
power tower system with a complex and high-precision secondary concentrator, as well as high-
temperature receiver materials [203]. Organic Flash Cycles (OFC) for CSP have also been
recommended for power generation with a medium-temperature heat source [358].
Another potential thermodynamic CSP cycle is the OFC which uses organic fluids, similar
to the ORC however with inherent advantages of a dry fluid as well as increased turbomachinery
performance from the higher molecular weight when compared with water. In an OFC all of the
heat is provided in the liquid phase, and the saturated liquid is throttled to a saturated vapor state
at lower pressure [203], which then passes through a turbine for power generation. This system
design reduces heating irreversibilities due to the temperature profile mismatch during
isothermal/isobaric phase change, allowing for closer matching of the source and sink temperature
profiles as in a single-phase heat exchanger. The OFC was previously shown to have higher
utilization efficiency than basic ammonia-water Rankine and transcritical CO2 cycles, favoring its
use as a bottoming cycle [203].

133
7. CSP POWER BLOCK HEAT EXCHANGE COMPONENTS
7.1 Heat Exchangers

Techno-economic analyses have shown a significant potential for cost reduction in CSP systems
through efficiency improvement of the power block, with particular focus to its stage components
[66, 307]. Heat exchangers are devices used to transfer enthalpy between two or more fluids
through a solid surface and a fluid, or between solid particles and a fluid. This occurs based on the
fluids operating at different temperatures, where traditionally there is no externally-applied heat
or work interactions. The portion of the surface that is in direct contact with both hot and cold
fluids, which transfers heat between them, is referred to as the primary or direct surface. To
increase the heat transfer area, supplemental attachments may be extended from the primary
surface to provide a secondary, or indirect surface. These extended surface elements are referred
to as fins. Thus, heat is conducted through the fin and convected (and/or radiated) from the fin
(through the surface area) to the surrounding fluid, or vice versa, depending on whether the fin is
being cooled or heated. As a result, the addition of fins to the primary surface reduces thermal
resistance on that side and thereby increases the total heat transfer from the surface for the same
temperature difference. Typically, in CSP applications, heat exchangers are used to recover or
reject heat, distill, concentrate, fractionate or control a process fluid.
Combustion and chemical reaction may take place within the exchanger, such as in boilers,
fired heaters, and fluidized-bed exchangers. Mechanical devices may be used in some exchangers
such as in scraped surface exchangers, agitated vessels, and stirred tank reactors. Heat transfer in
the separating wall of a recuperator generally takes place by conduction. If no phase change occurs
in any of the fluids within the exchanger, it is referred to as a sensible heat exchanger. Conversely,
a heat pipe heat exchanger not only acts as a separating wall, but also facilitates heat transfer by
phase change condensation and evaporation processes. In general, if the fluids are immiscible, the
separating wall may be eliminated, and the interface between the fluids replaces a heat transfer
surface, as in a direct-contact heat exchangers [359]. In his book Shah [359] provided
classifications of heat exchangers according to their transfer processes, heat transfer mechanisms,
construction type and flow arrangements. Fig. 7.1 provides a breakdown of these types, including
a classification for compact/non-compact heat exchangers which includes their heat transfer
surface area/volume ratio.

134
 a.

b.

c.

d.

e.
Figure 7.1. Heat exchanger classifications according to a. heat transfer process, b. construction, c.
flow arrangements, d. heat transfer mechanisms and e. surface compactness [359].
Overall, heat exchangers can be categorized according to their process function and whether they
are either indirect- and direct-contact types. In an indirect-contact heat exchanger, the fluid streams
remain separate where the heat transfers continuously through an impervious dividing wall or into
and out of a wall in a transient manner. Although a simultaneous flow of two (or more) fluids is
required within the exchanger, there is no direct mixing of the two (or more) fluids since each fluid
flows through separate fluid passages. Further categorization of these indirect/direct heat
exchangers according to their process functions can be seen in Fig. 7.2, as prescribed by Shah
[359].

135
 a.

b.

c.

d.

e.
Figure 7.2. Heat exchanger categorization according to a. process function, b. classification of
condensers and c. liquid-vapor exchangers [359].

Overall Shah and Sekulic [360] provide design criteria for facilitating successful heat exchanger
design:

1. The process requirements which accomplish thermal change on the streams within respective
allowable pressure drops, while retaining the capability to do this in the presence of fouling
until the next scheduled maintenance period.
2. The heat exchanger must withstand service conditions of the plant environment.

136
3. The exchanger must be maintainable, to facilitate a configuration that permits cleaning and
replacement of any component that is especially vulnerable to corrosion, erosion, vibration, or
aging.
4. The designer can consider advantages of a multishell arrangement with flexible piping and valve
designs, provided to allow one unit to be taken out of service for maintenance without
significant disturbance to the rest of the plant.
5. The heat exchanger should cost as little as possible provided that the above criteria are satisfied.
6. Limitations on the heat exchanger length, diameter, weight, and/or tube specifications due to
site requirements, lifting and servicing capabilities must be all taken into consideration in the
design.

Two general heat exchanger designs used in analysis are parallel flow and cross-flow. For parallel
flow heat exchangers, as shown in Fig. 7.3, the temperature difference between hot and cold fluids
varies with position within the heat exchanger as: ΔT = TH-TC. The temperature variations in heat
exchanger fluid-to-fluid heat transfer processes depend on flow path arrangement, where typical
configurations are: A. Parallel flow, B. Counterflow and C. Crossflow. In a parallel-flow
configuration the hot and cold fluids enter, flow and exit in the same directions versus a
counterflow heat exchanger where the two fluids flow in opposing directions, Fig. 7.4. These two
configurations are used extensively in two of the simplest heat exchangers: a concentric tube or
double-pipe configuration.

a. b.
Figure 7.3. Fluid a. heat transfer and b. temperature distributions for a parallel flow heat
exchanger.

From the first law of thermodynamics for an open system with steady state and flow conditions,
as well as with negligible changes in potential and kinetic energy, the heat transfer rate to the fluid
based on an enthalpy change from inlet to outlet locations A and B respectively is provided by:

𝑄̇ = 𝑚̇(ℎ𝐴 − ℎ𝐵 ) (7.1)

For two streams passing through a heat exchanger, one stream enters with a higher temperature, H
and one with a cooler temperature C, where if the fluids do not undergo a phase change process
and have constant specific heat values with dh = CpdT, then the following equations apply to the
system:
𝑄̇𝐻 = (𝑚̇𝐶𝑝 )𝐻 (ℎ𝐻,𝐴 − ℎ𝐻,𝐵 ) (7.2)

137
 𝑄̇𝐶 = (𝑚̇𝐶𝑝 )𝐶 (ℎ𝐶,𝐴 − ℎ𝐶,𝐵 ) (7.3)

a. b.
Figure 7.4. Fluid a. heat transfer and b. temperature distributions for a counter flow heat
exchanger.

Another common configuration is cross flow where heat exchange fluids flow perpendicular to
each other, which is typical of finned tubular heat exchangers that are considered to have mixed
or unmixed motion over the tubes. As shown in Fig. 7.5 if the fins inhibit motion in a direction
that is traverse to the main-flow direction then it is considered to be mixed where turbulence and
relatively higher heat transfer coefficients are possible. However, if the fluid flow passes through
the channel without significant obstruction then the flow is considered unmixed where temperature
variations are primarily in the main-flow direction [12].

Figure 7.5. Cross-flow heat exchanger with mixed and unmixed fluid transport [12].
As discussed in section 2.3 one method for computing the temperature difference between two
fluid streams is by computation of the log-mean temperature difference (LMTD), or ∆𝑇𝑙𝑚 . This
parameter can then be used to determine the rate of heat transfer by Eqn. 7.4, where U is the overall
heat transfer coefficient, used in place of a single convection coefficient h.
𝑄̇ = 𝑈𝐴Δ𝑇𝑙𝑚 (7.4)
When the inlet and outlet fluid temperatures are unknown however the LMTD method becomes
challenging as iterative procedures are then required [12]. Accordingly, the performance of a heat
exchanger may be also expressed with respect to the number of heat transfer units (NTU) of which
the heat exchanger is capable, where an NTU is defined by:

138
 𝑈𝐴
𝑁𝑇𝑈 = (7.5)
𝐶̇𝑚𝑖𝑛

A commonly used metric for evaluating heat exchanger performance is the effectiveness, defined
as the ratio of the actual heat transfer rate to the maximum heat transfer rate possible:
𝑄̇ 𝐶 (𝑇𝐻,𝑖𝑛 −𝑇𝐻,𝑜𝑢𝑡 ) 𝐶 (𝑇𝐶,𝑜𝑢𝑡 −𝑇𝐻,𝑖𝑛 )
𝜀 = 𝑄̇ = 𝐶𝐻 = 𝐶𝐶 (7.6)
𝑚𝑎𝑥 𝑚𝑖𝑛 (𝑇𝐻,𝑖𝑛 −𝑇𝐶,𝑖𝑛 ) 𝑚𝑖𝑛 (𝑇𝐻,𝑖𝑛 −𝑇𝐶,𝑖𝑛 )

It is then possible to express ε as a function of NTU, the heat capacity rate ratio C* and the flow
arrangement for any heat exchanger:
𝜀 = 𝑓(𝑁𝑇𝑈, 𝐶 ∗ , 𝐹𝑙𝑜𝑤 𝐴𝑟𝑟𝑎𝑛𝑔𝑒𝑚𝑒𝑛𝑡) (7.7)
where dimensionless heat capacity ratio and NTU are respectively defined as C*=Cmin/Cmax, Eqn.
7.5. It should be noted also that Cmin/Cmax = CC/CH = CH/ CC. For different heat exchanger flow
arrangements, Incropera and DeWitt provide various relations, Table 7.1 for effectiveness [12]
where the nature of the resulting temperature profiles depends on mass-heat capacity (𝑚̇𝐶𝑝 ) ratios
of the fluids. In this table NTU equations are also provided for simplicity.
As illustrated by Shah [359] in Fig. 7.6, these classifications provide a variety of heat
exchanger designs with thin-walled circular tubes and varying surface area densities. Also included
are different heat exchanger surface types with respective β values for plate, plate-fin and
regenerators:
𝐴
𝛽 = 𝑉𝐻 𝑜𝑟 𝐶 (7.8)
𝐻 𝑜𝑟 𝐶

where the porosity between plates or of a matrix surface is approximated as 0.8333, which provides
the relation 𝛽 = 3333⁄𝐷ℎ . For tube-fin and shell-and-tube heat exchangers, β is defined as:
𝐴𝐻 𝑜𝑟 𝐶
𝛽= (7.9)
𝑉𝑡𝑜𝑡𝑎𝑙

where A and V are the heat transfer area and exchanger volume respectively, while H and C denote
either hot or cold fluid sides. According to Shah [359]. Compact surface heat exchangers can gain
specified heat exchanger performance by q/Tm, within acceptably low mass and box volume
constraints, where this metric can be expressed in terms of the overall heat transfer coefficient as:
𝑞
= 𝑈𝐴 = 𝑈𝛽𝑉 (7.10)
Δ𝑇𝑚

where ΔTm is the true mean temperature difference, where a high β value minimizes exchanger
volume for a specified value of q/Tm. As compact surfaces can achieve structural stability and
strength with thinner-gauge material, the gain in a lower exchanger mass is even more pronounced
than the gain in a smaller volume [359].

139
 Table 7.1. Heat exchanger effectiveness correlations [12].
Flow Arrangement Relation
Concentric Tube
1−𝑒𝑥𝑝[−𝑁𝑇𝑈(1+𝐶𝑟 )]
𝜀= 1+𝐶𝑟
(7.11)
Parallel Flow
𝑙𝑛[1−𝜀(1−𝐶𝑟 )]
𝑁𝑇𝑈 = − 1−𝐶𝑟
(7.12)

1−𝑒𝑥𝑝[−𝑁𝑇𝑈(1+𝐶𝑟 )]
𝜀 = 1−𝐶 𝑓𝑜𝑟 𝐶𝑟 < 1 (7.13)
𝑟 𝑒𝑥𝑝1−𝑒𝑥𝑝[−𝑁𝑇𝑈(1+𝐶𝑟 )]
𝑁𝑇𝑈
𝜀 = 1+𝑁𝑇𝑈 𝑓𝑜𝑟 𝐶𝑟 = 1 (7.14)
Counter Flow
1 𝜀−1
𝑁𝑇𝑈 = 𝑙𝑛 ( ) 𝑓𝑜𝑟 𝐶𝑟 < 1 (7.15)
𝐶𝑟 −1 𝜀𝐶𝑟 −1
𝜀
𝑁𝑇𝑈 = 𝑓𝑜𝑟 𝐶𝑟 = 1 (7.16)
1−𝜀

Shell & Tube

1⁄2 −1
1+𝑒𝑥𝑝[−𝑁𝑇𝑈(1+𝐶𝑟2 ) ]
𝜀1 = 2 {1 + 𝐶𝑟 + (1 + 𝐶𝑟2 )1⁄2 1⁄2 } (7.17)
1−𝑒𝑥𝑝[−𝑁𝑇𝑈(1+𝐶𝑟2 ) ]

One Shell Pass 𝐸−1

𝑁𝑇𝑈 = −(1 + 𝐶𝑟2 )−1⁄2 𝑙𝑛 (𝐸+1)
2⁄𝜀1 −(1−𝐶𝑟 ) (7.18)
𝐸= 1⁄2
(1+𝐶𝑟2 )
−1
1−𝜀1 𝐶𝑟 𝑛 1−𝜀1 𝐶𝑟 𝑛
𝜀 = [( 1−𝜀1
) − 1] [( 1−𝜀1
) − 𝐶𝑟 ] (7.19)
n shell passes
𝐹−1 𝜀𝐶𝑟 −1 1⁄𝑛
Use Eqns. 7.15 and 7.16 with: 𝜀1 = ,𝐹 =( ) (7.20)
𝐹−𝐶𝑟 𝜀−1

Cross Flow (Single Pass)

1
Both Fluids are Unmixed: 𝜀 − 1 − 𝑒𝑥𝑝 [(𝐶 ) 𝑁𝑇𝑈 0.22 {𝑒𝑥𝑝[−𝐶𝑟 (𝑁𝑇𝑈)0.78 ] − 1}] (7.21)
𝑟

1
CMax (Mixed) 𝜀 = (𝐶 ) {1 − 𝑒𝑥𝑝[−𝐶𝑟 (1 − 𝑒𝑥𝑝(−𝑁𝑇𝑈))]} (7.22)
𝑟

CMin (Mixed) 1
𝑁𝑇𝑈 = −𝑙𝑛 [1 + (𝐶 ) 𝑙𝑛(1 − 𝜀𝐶𝑟 )] (7.23)
𝑟

CMax (Unmixed) 𝜀 = 1 − 𝑒𝑥𝑝(−𝐶𝑟−1 {1 − 𝑒𝑥𝑝[−𝐶𝑟 (𝑁𝑇𝑈)]}) (7.24)

1
CMin (Unmixed) 𝑁𝑇𝑈 = − (𝐶 ) 𝑙𝑛[𝐶𝑟 𝑙𝑛(1 − 𝜀) + 1] (7.25)
𝑟

𝜀 = 1 − exp(−𝑁𝑇𝑈) (7.26)
All Exchangers (Cr = 0)
𝑁𝑇𝑈 = −𝑙𝑛(1 − 𝜀) (7.27)

140
 Figure 7.6. Heat transfer surface area density spectrum of heat exchanger surfaces [359].

For gas-to-fluid exchangers, the heat transfer coefficient h for gases is generally one or more orders
of magnitude lower than for liquids such as water or oil. To minimize size, weight and overall
costs, thermal conductances (ha products) for both sides of the exchanger should be in value, where
the heat transfer surface on the gas-side needs to have a larger surface area and be more compact
to facilitate this [359].

7.1.1 Gasketed-Plate Heat Exchangers

Gasketed plate heat exchangers utilize metal plates to transfer heat between fluids where its
advantage over other heat exchanger designs is that the fluids are exposed to a much larger surface
area since the fluids spread out over the plates. This facilitates the transfer of heat, which greatly
increases the rate of temperature change. Plate heat exchangers fundamentally differ where their
design replaces pipes that would pass through a chamber for heat transfer. Instead there are two
alternating chambers, separated at their largest surface by a corrugated metal plate. These heat
exchangers are highly configurable due to the variety of designs available for the plates and
arrangements that are possible to suit a variety of applications. Unlike tubular heat exchangers for
which design data and methods are easily available [360], gasketed-plate heat exchanger designs
continue to be proprietary in nature where attempts have been made to develop heat transfer and
pressure drop correlations. However, most of these cannot be generalized to give a high degree of
prediction accuracy [360]. For these heat exchangers, the fluids are much closer to countercurrent
flow than in shell-and-tube heat exchangers. Further, another permutation of this design is a
gasketed plate-and-frame heat exchanger, which can provide efficient heat transfer in compact

141
equipment with a small footprint. These devices have been shown to have flexible design and are
easy to service and maintain [361]. Stainless steel is commonly used for the plates due to its ability
to withstand high temperatures, its high strength and corrosion resistance [362]. The plates are
often spaced by rubber sealing gaskets which are cemented into a section around the edge of the
plates. The plates are pressed to form troughs at right angles to the direction of flow of the liquid
which runs through the channels in the heat exchanger. These troughs are arranged so that they
interlink with the other plates which forms the channel with gaps of approximately 1.3–1.5 mm
between the plates [363]. The plates are compressed together in a rigid frame to form an
arrangement of parallel flow channels with alternating hot and cold fluids. The plates, with their
large surface area, can facilitate high heat transfer rates. Making each chamber thin ensures that
the majority of the volume of liquid contacts the plate, again aiding exchange. The troughs also
create and maintain a turbulent flow, even at low flow rates, to maximize heat transfer coefficients
[363].
Compared to shell and tube heat exchangers, the temperature approach in a plate heat
exchanger may be as low as 1 °C, where shell and tube heat exchangers require an approach of 5
°C or more. Therefore, for the same amount of heat exchange, plate heat exchanger size can be
smaller due to the large heat transfer area afforded by the plates. Increase and reduction of the heat
transfer area is relatively simple with plate heat-exchangers through the addition or removal of
plates from the stack [363] where heat transfer enhancement will strongly depend on the Chevron
inclination angle β, relative to flow direction, as well as the friction factor that increases with β.
The Chevron angle typically has a range between 25ᵒ-65ᵒ [360], and as shown in Fig. 7.7,
adjustment of the corrugation pitch and depth increases the flat or projected plate area.

Figure 7.7. Gasketed-plate with chevron design grooves [360].

The increase of the developed length with respect to the projected length, can be described by an
enlargement factor 𝜙 which can vary between 1.1 and 1.25 [422].

142
 𝑑𝑒𝑣𝑒𝑙𝑜𝑝𝑒𝑑 𝑙𝑒𝑛𝑔𝑡ℎ
𝜙= (7.28)
𝑝𝑟𝑜𝑗𝑒𝑐𝑡𝑒𝑑 𝑙𝑒𝑛𝑔𝑡ℎ

Another definition of 𝜙 is the ratio of the actual effective area to the projected plate area where the
projected area can be approximated by 𝐿𝑃 ∙ 𝐿𝑤 . The determination of other geometrical performance
parameters, such as the mean channel flow gap and channel flow area can be found by Kakaç et al.
[360]. For gasketed-plate exchangers, heat transfer enhancement is determined based on the Chevron
inclination angle β, relative to flow direction, which influences friction factor. However, performance
of a Chevron plate also depends on the surface enlargement factor 𝜙, corrugation profile, and
temperature dependent physical properties, such as viscosity [360]. Comprehensive analyses of various
gasketed-plate heat transfer correlations are provided and compared by Manglik [423] and Kakaç et al.
[360] where discrepancies originate from differences in plate surface geometries such as: surface
enlargement factor 𝜙, metal-to-metal contact pitch, wavelength, amplitude, and profile or shape of the
surface corrugation. A well accepted correlation for Chevron plates, based on an extensive data by
Muley and Manglik [424-426] provide a heat transfer coefficient based on β and Re.

2ℎ𝑏 𝛽 0.38 𝜇 0.14

𝑁𝑢 = 𝑘
= 0.44 (30) 𝑅𝑒 0.5 𝑃𝑟 1⁄3 (𝜇 𝑏 )
𝑤
0.2 } 𝑓𝑜𝑟 𝑅𝑒 ≤ 400 (7.29)
𝛽 0.83 30.2 5 6.28 5
𝑓=( ) [( 𝑅𝑒 ) + (𝑅𝑒 0.5) ]
30

𝜇 0.14
𝑁𝑢 = [0.2668 − 0.006967𝛽 + 7.244 × 10−5 𝛽2 ]𝑅𝑒 𝜁 𝑃𝑟 1⁄3 (𝜇 𝑏 )
𝑤
𝜁 = 0.728 + 0.0543𝑠𝑖𝑛[(2𝜋𝛽⁄90) + 3.7] } 𝑓𝑜𝑟 𝑅𝑒 ≤ 800 (7.30)
𝑓 = [2.917 − 0.1227𝛽 + 2.016 × 10−3 𝛽2 ]𝑅𝑒 −{0.2+0.0577𝑠𝑖𝑛[2𝜋𝛽⁄90+2.1]}

The investigators also developed prediction correlations for flat-plate channels, Eqns. 7.31 and 7.32.

𝑁𝑢 = 1.849(𝐿⁄𝑑𝑒 )−1⁄3 (𝑅𝑒𝑃𝑟)1⁄3 (𝜇𝑏 ⁄𝜇𝑤 )0.14 𝑓𝑜𝑟 ≤ 2000

(7.31)
𝑁𝑢 = 0.023𝑅𝑒 0.8 𝑃𝑟 1⁄3 (𝜇𝑏 ⁄𝜇𝑤 )0.14 𝑓𝑜𝑟 > 4000

𝑓 = 24⁄𝑅𝑒 𝑓𝑜𝑟 𝑅𝑒 ≤ 2000

(7.32)
𝑓 = 0.1268𝑅𝑒 −0.3 𝑓𝑜𝑟 𝑅𝑒 > 2000

Muley and Manglik compared these correlations with those of Kakaç, with a Chevron angle of 𝛽 =
1.29 with a channel aspect ratio of 2𝑏⁄𝑃𝑐 , where they found with respect to Re, the Chevron plates
were able to produce up to five times higher Nu values than flat-plate channels, however with a
considerably higher pressure drop. In general, various correlations are available for plate heat
exchangers for various fluids depending on flow characteristics and plate geometries [360].
Finally, designed thermal mixing between stacked pates can provide additional flexibility in
utilization of the available pressure drop without excessive surface area, and with fewer standard plate
patterns [360]. The level of mixing is sensitive to 𝛽 where a pack of plates may be composed of all
high angles, low angles, and may be arranged alternately to provide an intermediate level of
performance.

143
 a. b. c.

Figure 7.8. Thermal mixing configurations for a. low-low β angles, b. mixed β angles and c. high-high
β angles [360].

7.1.2 Shell and Tube Heat Exchangers

In contrast to compact heat exchangers, general industry-standard shell and-tube exchangers have
a surface area density of less than 100 m2/m3 on one side with plain tubes, and a surface area
density two to three times greater with high-fin-density, low-finned tubing [359]. Additionally, for
water/water applications [365], a typical plate heat exchanger has approximately twice the average
heat transfer coefficient h on one fluid side for a shell and-tube exchanger. Plate-fin, tube-fin, and
rotary regenerators are examples of compact heat exchangers for gas flow on one, or both fluid
sides. Additionally, gasketed, welded, brazed-plate and printed-circuit heat exchangers are
examples of compact heat exchangers for liquid flows. Basic flow arrangements of two-fluid
compact heat exchangers are single-pass crossflow, counterflow, and multipass cross-counterflow
for non-compact heat exchangers. The last two flow arrangements for compact or non-compact
heat exchangers can yield very high exchanger effectiveness values or very small temperature
differences between the fluid streams [365].
Shell-and-tube exchangers are traditionally custom designed for virtually any capacity and
operating condition, from high vacuums to ultrahigh pressure systems; from cryogenics to high
temperatures, and for any temperature/pressure differences between fluids limited only by
respective structural materials. They can be designed for special operating conditions: vibration,
highly viscous fluids, erosion, corrosion, toxicity, radioactivity, multicomponent mixtures, where
are made from a variety of metal and nonmetal materials, and in surface areas from less than 0.1
to 100,000 m2 [364]. Generally, shell-and-tube heat exchangers are an order of magnitude less
surface area per unit volume than the compact exchangers, where they also require considerable
space, weight, support structure, and footprint [169]. Fig. 7.9 by Kakaç et al. [364] provides a
graphical illustration for a shell-and-tube heat exchanger that comprises tube bundles, baffles and
the shell. Stream B is the main cross flow stream which passes from one window, across the cross
flow section, and out through the opposite window. Additionally, there are four other streams
because of the mechanical clearances required in a shell-and-tube heat exchanger. One is the A
stream that leaks through the clearance between the tubes and the baffle, from one baffle

144
compartment to the next. There is also a C stream which is the bundle bypass stream which flows
around the tube bundle between the outermost tubes and the inside of the shell. The E stream flows
through a clearance between the baffles and the inside diameter of the shell. The F stream flows
through any channels within the tube bundle, facilitated by the provision of pass dividers in the
exchanger header [364].

Figure 7.9. Generalized shell-in-tube heat exchanger, which portrays leaking paths for flow bypassing
the tube matrix through the baffle clearances between the tube matrix and the shell [364].

For heat transfer analysis, two prominent methods used in industry are the Bell-Delaware method
[366] which is most-widely used, and the Kern method [367], which offers simplicity though with
some notable restrictions and limitations [364]. Here the Kern method will be discussed, which
was developed as an attempt to correlate data for standard exchangers by a simple equation
analogous to equations for flow in tubes. However, this method is restricted to a fixed baffle and
cannot adequately account for baffle-to-shell and tube-to-baffle leakages [364]. For this method,
thermal analysis begins by calculating the mass flow rate as a function of the number of tubes Nt
and by considering the cross sectional area 𝐴𝑐 = 𝜋𝐷𝑖2 ⁄4 where Di is the inner tube diameter.

𝑚̇𝑡 = 𝜌𝑡 𝑢𝑡 𝐴𝑐 𝑁𝑡 (7.33)

For laminar flow the Sieder and Tate correlation [431] can be used where 𝑅𝑒𝑡 and 𝑃𝑟𝑡 are defined
for the tube-side fluid and 𝐿𝑐 is the tube characteristic length.

𝑅𝑒𝑡 𝑃𝑟𝑡 𝐷𝑖 1⁄3 0.48 < 𝑅𝑒𝑡 < 16700

𝑁𝑢𝑡 = 1.86 ( ) 𝑓𝑜𝑟 (7.34)
𝐿𝑐 (𝑅𝑒𝑡 𝑃𝑟𝑡 𝐷𝑖 ⁄𝐿𝑐 )1⁄3 > 2

For turbulent flow, the Nusselt number can be analyzed by the Petukhav-Kirillov correlation [430].

(𝑓⁄2)𝑅𝑒𝑡 𝑃𝑟𝑡 104 < 𝑅𝑒𝑡 < 5 × 106

𝑁𝑢𝑡 = 2⁄3 𝑓𝑜𝑟 (7.35)
1.07+12.7(𝑓⁄2)1⁄2 (𝑃𝑟𝑡 −1) 0.5 < 𝑃𝑟𝑡 < 2000

where (𝑓⁄2) is assumed to be 0.0029 and the friction factor determined by:

𝑓𝑡 = (1.58𝑙𝑛𝑅𝑒𝑡 − 3.28)−2 (7.36)

145
For the shell-side, the number of tubes required is determined by taking the quotient of the shell
circle by the projected area of the tube layout, Eqn. 7.37:

𝜋𝐷𝑠2
𝑁𝑡 = 𝐶𝑇𝑃 4𝐴 (7.37)
1𝑡𝑢𝑏𝑒

From Eqn. 7.37 the projected area 𝐴1𝑡𝑢𝑏𝑒 corresponds to the area of one tube, which can be
expressed as:
𝑓𝑜𝑟 90° 𝑎𝑛𝑑 45° → 𝐶𝐿 = 1.0
𝐴1𝑡𝑢𝑏𝑒 = (𝐶𝐿)𝑃𝑡2 (7.38)
𝑓𝑜𝑟 30° 𝑎𝑛𝑑 60° → 𝐶𝐿 = 0.87

Ds is the shell inner diameter and CTP is the tube count calculation constant that accounts for
incomplete coverage of the shell diameter by tubes, which for different number of tube passages
can be provided by [364]:
𝑜𝑛𝑒 − 𝑡𝑢𝑏𝑒 𝑝𝑎𝑠𝑠 → 𝐶𝑇𝑃 = 0.93
𝑡𝑤𝑜 − 𝑡𝑢𝑏𝑒 𝑝𝑎𝑠𝑠 → 𝐶𝑇𝑃 = 0.90
𝑡ℎ𝑟𝑒𝑒 − 𝑡𝑢𝑏𝑒 𝑝𝑎𝑠𝑠 → 𝐶𝑇𝑃 = 0.85

The shell equivalent diameter is determined along the long axes of the shell as four times the net
flow area as layout on the tube sheet divided by the wetted perimeter [364]:

4×𝑓𝑟𝑒𝑒−𝑓𝑙𝑜𝑤 𝑎𝑟𝑒𝑎
𝐷𝑒 = (7.39)
𝑤𝑒𝑡𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟

As shown in Fig. 7.10 there are two general configurations for the flow over the tubes, where for
the square pitch and triangular pitch, the perimeter is determined by the pitch size PT which is the
outside tube diameter:
𝜋𝐷2
4(𝑃𝑡2 − 𝑜)
4
𝑆𝑞𝑢𝑎𝑟𝑒 𝑃𝑖𝑡𝑐ℎ: 𝐷𝑒 = 𝜋𝐷𝑜
2
𝑃 √3 𝜋𝐷 2 (7.40)
4( 𝑡 − 𝑜 )
4 8
𝑇𝑟𝑖𝑎𝑛𝑔𝑢𝑙𝑎𝑟 𝑃𝑖𝑡𝑐ℎ: 𝐷𝑒 = 𝜋𝑑𝑜 ⁄2

a. b.

Figure 7.10. Tube layouts for a. square-pitch and b. triangular-pitch [364].

To determine the shell-side heat transfer coefficient, the correlation by McAdams [428], Eqn. 7.41
is provided for the shell-side.

ℎ𝑜 𝐷𝑒 𝐷 𝑚̇ 0.55 𝐶 𝜇 1⁄3 𝜇 0.14

𝑝
𝑒 𝑠
= 0.36 ( 𝜇𝐴 ) ( ) (𝜇 𝑏 ) 𝑓𝑜𝑟 2 × 103 < 𝑅𝑒𝑠 < 1 × 106 (7.41)
𝑘 𝑠 𝑘 𝑤

146
 ṁ D
where: Res = ( A s ) μe (7.42)
s s

and De is the equivalent diameter of the shell side. As prescribed by [364] the pressure drop
encountered by the fluid making Np passes through the heat exchanger is a multiple of the kinetic
energy of the flow:
𝐿𝑁𝑝 𝜌𝑡 𝑢𝑡2
Δ𝑃𝑡 = (4𝑓𝑡 + 4𝑁𝑝 ) (7.43)
𝐷𝑖 2

where 𝑓𝑡 is determined from Eqn. 7.36, and the second term represents the additional pressure drop
introduced by the change of direction in the respective passages [364]. For the shell-side, the fluid
experiences a pressure drop as it passes through the exchanger, over the tubes and around the
baffles. If the shell fluid nozzles are on the same side of the heat exchanger, then the shell-side
fluid can make an even number of tube bundle crossings, which is therefore dependent on pressure
drop [364]. From experiments, the pressure drop experienced by the shell-side fluid is determined
by Eqn. 7.44 [429].
𝑓𝑠 𝐺𝑠2 (𝑁𝑏 +1)𝐷𝑠
Δ𝑃𝑠 = (7.44)
2𝜌𝑠 𝐷𝑒 𝜙𝑠

where Nb is the number of baffles, Ds is the inner shell diameter and:

𝐺𝑠 𝐷𝑒
𝑓𝑠 = 𝑒 0.576−0.19𝑙𝑛𝑅𝑒𝑠 𝑓𝑜𝑟 400 < 𝑅𝑒𝑠 = ≤ 1 × 106 (7.45)
𝜇

𝜙𝑠 = (𝜇𝑏 ⁄𝜇𝑤 )0.14 (7.46)

7.1.3 Compact Heat Exchangers

As shown in Fig. 7.6 a gas-to-fluid heat exchanger is referred to as a compact heat exchanger
(CHE) if it incorporates a heat transfer surface having a surface area density above approximately
700 m2/m3 on at least one of the fluid sides, which usually has gas flow. It is referred to as a laminar
flow heat exchanger if the surface area density is above about 3000 m2/m3, and as a micro heat
exchanger if the surface area density is above approximately 10,000 m2/m3 [169]. A liquid/two-
phase heat exchanger is referred to as a CHE if the surface area density on any one fluid side is
above about 400 m2/m3 [169]. A typical process industry shell-and-tube exchanger has a surface
area density of less than 100 m2/m3 on one fluid side with plain tubes, and two to three times that
with the high-fin-density low-finned tubing. Plate-fin, tube-fin, and rotary regenerators are
examples of CHEs for gas flows on one or both fluid sides, and gasketed and welded plate heat
exchangers are examples of compact heat exchangers for liquid flows [169]. A detailed overview
of heat exchanger types can be found by Hesselgreaves [368]. For compact extended surface heat
exchangers, important design and operating considerations suggested by Kreith [169] are:

1. At least one of the HTFs is a gas or specific liquid with low h values.
2. Fluids must be clean and relatively noncorrosive because of small hydraulic diameter Dh flow
passages with generally no easy techniques for cleaning.
3. The fluid pumping power (pressure drop) design constraint is often equally as important as the
heat transfer rate with respect to performance.

147
4. Operating pressures and temperatures are somewhat limited compared with shell-and-tube heat
exchangers as a result of the joining of the fins to plates or tubes such as brazing, mechanical
expansion, etc.
5. With respect to the use of highly compact surfaces, the resultant shape of a gas-to-fluid
exchanger is one having a large frontal area and a short flow length (the header design of a CHE
is thus important for a uniform flow distribution among the very large number of small flow
passages).

Some advantages of plate-fin exchangers over conventional shell-and-tube exchangers are that
CHEs, generally fabricated from thin metallic plates can yield a large heat transfer surface area
per unit volume, typically up to ten times greater than the 50-100 m2/m3 density provided by a
shell and- tube exchanger for general process application, from 1000 to 6000 m2/m3 for highly
compact gas side surfaces [169].
With regards to heat exchange limitations, small flow passages have two effects, a tendency
for facilitating laminar flow within the channels, and a high pressure drop. Laminar flow is
associated with low heat transfer coefficients, and therefore the efficiency is necessarily improved
by various heat transfer enhancement techniques, which have facilitated a variety of CHE designs.
A CHE provides a tighter temperature control and can provide rapid heating or cooling of a process
stream, thus improving the product quality. CHEs are characterized by a large heat transfer surface
area per unit volume of exchanger, resulting in reduced space, weight, support structure, energy
requirements and cost, as well as improved process design, processing conditions and low fluid
inventory. Its high-surface area density design is able to facilitate higher heat transfer coefficients,
which make it inherently a smaller less-bulky device compared to other heat exchangers.
Traditionally, this type of heat exchangers can be classified according to respective surface area
densities and hydraulic diameters, with approximate values given in Table. 7.2 [365].

Table 7.2. Compact heat exchanger classification [365].

For CO2 applications, compact heat exchangers such as the RANOTOR as described by Chen
[331], are designed to be counter current with laminar flow at the heat exchanger gas side, and
composed of concentric conical plates, which are comprised of several tubes arranged in parallel
as shown in Fig. 7.11. This design has the advantages of high power densities, a low pressure drop
and flexible installation [331].

148
 a. b.
Figure 7.11. RANOTOR compact heat exchanger [331].

From an O&M point of view, compact heat exchangers are used in a variety of configurations,
such as high temperature (<850°C), high pressure applications (>200 bars), and in applications
that encompass a moderate amount of fouling. However, a large majority of reported applications
do not involve both aforementioned high temperature and pressure conditions simultaneously. For
many compact exchangers except for plate heat exchangers, fouling is a significant challenge. For
an exchanger with a large frontal area, flow maldistribution could be another potential issue.
Additionally, because of short transient times, a careful design of controls is required for startup
of compact heat exchangers compared with shell-and-tube exchangers [331].

7.1.4. Plate Heat Exchangers

There are a variety of locations where CHEs can be employed throughout a CSP system from
receivers [369] to the power block [371]. One type of CHE is a plate heat exchanger (PHE), Fig.
7.12 [370], which consists of a series of corrugated plates supported by a rigid frame. The
corrugations on adjacent plates contact or cross each other forming highly interrupted and tortuous
channels. Sealing between streams is accomplished by gaskets. In addition, multi-pass
configurations can be accommodated by blanking plates within the stack. Various system
specifications can be matched by adding or removing plates or inclusion of plates with differing
patterns.

Figure 7.12. Generalized plate heat exchanger configuration [370]

149
The plate-fin exchangers can accommodate multiple fluid streams in one exchanger with proper
manifolding, which allows process integration and cost effective compact design solutions.
Fouling can be a major operational challenge with compact heat exchangers (except for plate-and
frame heat exchangers), particularly those having a variety of fin geometries or very fine circular
or noncircular flow passages that cannot be cleaned mechanically. Chemical cleaning may be
possible, where thermal baking and subsequent rinsing is possible for small-size units. Therefore,
extended surface compact heat exchangers may not be optimal for use in heavy fouling
applications [169]. One advantage of PHEs is the greatly reduced space requirements.
The space required for a PHE can be 30–50% of a shell-and-tube heat exchanger for a
given heat duty, [372]. Additionally, for the same effective heat transfer area, the weight and
volume of PHEs are approximately only 30% and 20%, respectively, of shell-and-tube heat
exchangers [372]. They can have 50% less volume than a fined tube heat exchanger, and 60% less
than a serpentine design for the same thermal performance [373]. Choice of gasket materials is
also critical to reliable operation of PHEs [375], where gaskets are generally comprised of a variety
of elastic and formable materials, such as rubber [372]. Gasket materials restrict the utility of PHEs
in highly corrosive applications which however limit the maximum operating temperature to avoid
higher costs. The thermal conductivity of the plate is an important consideration for thermal–
hydraulic design of a PHE, therefore plate materials with higher thermal conductivity are preferred.
A wide variety of plate materials are available such as those suggested by Li [369]: Stainless steel
(types 304, 316, 317, 304L, 316L, 317Ti), Alloy AL6XN, Alloy 904L, Alloy 27- 7MO, Alloy 254
SMO; Nickel 200, Alloy G-30, Alloy B-2, Alloy C-22, Alloy C-276, Alloy C-2000, Alloy 33;
Titanium (Gr. 1), Titanium palladium (Gr. 7&11), tantalum. Of these, stainless steel is the most
commonly used for plates due to its ability for enduring high temperatures, as well as its strength,
and corrosion resistance [372, 375].

7.1.5. Plate-Fin Exchangers

A plate-fin heat exchanger (PFHE) is a CHE consisting of alternating layers of corrugated sheets
(fins), separated by parting sheets, and enclosed at the edges by side bars to create a series of finned
chambers.

Figure 7.13. Plate-fin heat exchanger qualitative illustration [367].

As illustrated in Fig. 7.13, the fins and the parting sheets are assembled by brazing in a vacuum
furnace to become a single rigid core, where this heat exchanger design can be comprised of one
or more cores. Performance optimization is generally facilitated based on the number of plate and

150
fin layers, the size of the plate and fin, the height of the fin and the type of fin [377, 378]. During
PFHE operation, heat is transferred from the hot fluid through the fin interface to the separator
plate and through the next set of fins into the adjacent cold fluid. The fins have two functions: 1.
first, to act as the secondary (or indirect, extended) heat transfer surface to increase the effective
heat transfer surface area and increase the total heat transfer from the surface, and 2. to withstand
the design pressure at the design temperature as a structural component [369]. Appendix A.1
provides general information on common application for each type of fin. In a PFHE, fins can be
easily rearranged, which allow the device to operate in cross-flow, counter-flow, cross-
counterflow flow. A simple cross-flow layout is generally suitable for low or moderate duties,
which is nominal when one side consists of a low pressure gas. For heavier duties, the counter-
flow pattern is generally a more productive solution, which also can be more efficient for most
lower-temperature applications.
For plain fins, heat transfer enhancement is generally achieved due to an increased area
density, rather than any increase in heat transfer coefficient [369]. Herringbone fins can enhance
heat transfer by promoting mixing due to generated complex secondary flows and boundary layer
separation. Serrated and perforated fins can be categorized as interrupted fins, which can enhance
heat transfer through two mechanisms. First, surface interruption can prevent continuous growth
of the thermal boundary layer by periodically interrupting it [379]. Secondly, the thicker thermal
boundary layer in continuous fins offer higher thermal resistance to heat transfer, and is maintained
thin and the resistance to heat transfer is reduced. This thermal enhancement mechanism occurs
even at low Reynolds numbers when the flow is steady and laminar. Above a critical Reynolds
number, the interrupted surfaces offer an additional mechanism of heat transfer enhancement by
inducing self-sustained oscillations in the flow in the form of shed vortices [369], which are
facilitated near the leading edge of the fins and travel downstream along the fin surface. Von
Karman vortices are also observed to form at the trailing edge of the fins and travel downstream
in the wake before encountering the next fin element. They act as large scale mixers and
continuously bring in fresh fluid from the free-stream on their downstream side towards the fin
surface [380, 381] and eject the fluid on their upstream side away from the fin surface. This
however facilitates an associated increased pressure drop and pumping power due to higher skin
friction associated with the hydrodynamic boundary layer resuming [369]. In addition, in the
unsteady flow regime, the time-dependent flow behavior associated with vortex shedding increases
frictional loss through the Stokes layer dissipation and form drag through Reynolds stresses [382].
The two independent mechanisms simultaneously influence both the overall heat transfer and the
pumping power requirement [369].
For PFHEs, heat transfer performance is based on fin efficiency, Eqn. 7.47, which is
determined by the surface area weighted average of finned section, 𝜂𝑓,𝑖 .

∑𝑛
𝑖=1 𝜂𝑓,𝑖 𝐴𝑓,𝑖
𝜂𝑓 = ∑𝑛
(7.47)
𝑖=1 𝐴𝑓,𝑖

According to Huang and Shah [383], Eqn. 7.47 is not valid in when the fin is thick, subject to
variable heat transfer coefficients or variable ambient fluid temperature, or when it has a
temperature depression at the fin base. However, for thin rectangular fins of constant cross-section,
Shah [169] provides fin efficiencies for various cross-sectional plate-fin and tube-fin geometries.

151
7.1.6. Printed Circuit Heat Exchanger

Printed-circuit heat exchangers (PCHE), Fig. 7.14, are a category of CHEs due to their high surface
density area (>2500 m2/m3) [384], where plate sizes can be as high as 1.2 m × 0.6 m [377]. PCHEs
are strong candidate heat exchangers for CSP receiver and power block loops as they can achieve
high power densities with far less material than comparable shell-and-tube heat exchangers [385].
The high temperatures within intermediate loops require heat exchangers that retain their strength
at high temperatures [386], where PCHE’s can be strong candidates. Their fluid passages are
approximately semicircular in cross-section, being typically 1.0–2.0 mm wide, 0.5–1.0 mm in
depth [387], with approximately 0.5–2.0 mm hydraulic diameters in size. PCHEs employ the same
manufacturing processes as for printed circuit boards which involves photo-chemically etched
grooves on a flat metal plate to form fluid passages. The etched-out plates are thereafter alternately
joined by diffusion bonding, which can result in a compact, strong, all-metal heat exchanger core.

Figure 7.14. Cross-sectional view of a microchannel PCHE [388].

The diffusion bonding process includes a thermal soaking period to allow grain growth, thereby
potentially eliminating the interface at the joints, which provides base-material strength and very
high pressure containment capability throughout the entire exchanger, in addition to corrosion
resistance [369]. Because of diffusion bonding, its expected lifetime exceeds that of most heat
exchangers based on a brazed structure [289]. The complete heat exchanger core, can be tuned by
welding together multiple blocks as required by a desired thermal duty (flow capacity) of the heat
exchanger. The most commonly employed flow configurations include counter-flow, cross-flow
or co-flow, where other advanced characteristics can be added to contribute to further improving
PCHE performance [369].
In addition to its wide operating range, PCHEs also have enhanced safety features as they
characteristically do not contain any gaskets or brazed material. Consequently, the risk of leaks or
fluid incompatibility is substantially reduced [390]. According to Sabharwall [391] the risk of
leaks in a PCHE is approximately two orders of magnitude lower than for any other heat exchanger
due to its continuous passage [391]. PCHEs have been rated to operate under a temperature range
from cryogenic to 900 °C, while withstanding pressures as high as 60 MPa [392]. The design of
PCHEs enable operation under high purity or corrosive streams, and are not susceptible to flow
induced oscillations and vibrations, like that experienced by conventional shell and tube heat
exchangers. According to HeatricTM [393], PCHEs require low maintenance if used in flows with
particulate size limited to 300 μm.
Characteristic continuous zigzag PCHE flow paths do not allow boundary layer growth
and encourage into turbulent flow. By enhancing heat transfer area and increasing local flow
velocity at channel bending points, this zigzag channel shape enhances thermal performance

152
compared with heat exchangers that have straight channels [394]. Compared to straight ducts,
wavy geometries provide little advantage at low Reynolds numbers, although maximum advantage
at transitional Reynolds numbers. Gschwind [395] described the important flow mechanisms
associated with wavy channels, where low Re < 200 values facilitate steady recirculation zones
that form in the troughs of the wavy passages, which do not enhance heat transfer. However, for
higher Reynolds numbers, the free shear layer becomes unstable where subsequent vortices are
advected downstream. Additionally, at higher Reynolds numbers, periodic shedding of transverse
vortices increases the Nusselt number with a considerable increase in the friction factor [369].
According to Urquiza et al. [386] although PCHE heat exchangers can achieve high
effectiveness, they can also be susceptible to very large thermal stresses that occur during flow
transients, which can be significantly higher than those predicted under steady state conditions.
Peak stresses occur at the feature scale where the fin roots and the flow channel geometry
concentrate stresses resulting from thermal expansion [386].

7.1.7. Spiral Heat Exchangers

A spiral heat exchanger (SHE) is characterized by a helical tube configuration as illustrated by

Fig. 7.15 where the circular heat exchanger has two long metal strips rolled together to form a pair
of concentric spiral channels of rectangular cross-section, one for each fluid [396].

Figure 7.15. Spiral heat exchanger [396].

SHE passages can either be smooth or corrugated to enhance turbulent mixing, and in some cases
studs are welded onto one side of each strip to fix the spacing between the plates, providing
improved mechanical strength and increased turbulence and subsequent heat transfer [396].
Alternating passage edges are sealed either by welding at each side of the channel [368] or by
providing a gasket at each end cover to obtain the following configurations of the two fluids: (1)
both in spiral counter-flow; (2) one in spiral flow, with the other in cross-flow across the spiral; or
(3) one in spiral flow, with the other in a combination of cross-flow and spiral flow. With
countercurrent SHEs, an LMTD correction factor of 1.0 does not have an LMTD penalty like other
exchangers [396].
In non-corrugated SHE devices, heat transfer coefficients are not as high as in PHEs, which
are designed for operation at temperatures up to 540 °C, where SHEs are conventionally designed
to operate up to 200 °C [369]. As with gasketed PHEs, the temperature limit is dependent on the
gasket material used to seal the gaps between the open channel ends and the end covers. The fluid
in a SHE is fully turbulent at a much lower velocity than in straight tube heat exchangers due to
the studs and the concentric shape of the flow passages. As fluid flows, the curvature of the

153
passages facilitates a centrifugal force, which also impose secondary flow effects (eddy currents
and vortices) which significantly enhance heat transfer [397]. SHE heat transfer characteristics can
be improved by adjusting channel dimensions, such as variation of the diameter, surface
modification or by adding roughness to the channel walls. SHE devices conventionally have
relatively large diameters and a channel spacing range between 5 mm and 25 mm (representing
hydraulic diameters of 10–50 mm), where engineers can leverage their high heat transfer surface
area densities, and low propensities for blockage by fouling. Picon-Nunez et al. [398] presented a
shortcut method for the sizing of spiral plate heat exchangers with single-phase processes, where
SHEs can also be made compact. The internal void volume is lower (< 60%) than a shell-and-tube
heat exchanger [365], which can facilitate a compact, space-saving construction that can be readily
integrated in any plant, reducing installation costs. The heat transfer surface traditionally ranges
from 0.05 m2 for refrigeration applications, up to approximately 500 m2 with a maximum shell
diameter of 1.8 m and sheet metal thickness between 1.8–4 mm for industrial processes [377]. The
surface area requirement is about 20% lower than that for shell-and-tube heat exchangers for the
same heat duty [365].

7.1.8. Ceramic Heat Exchangers

Ceramic heat exchangers (CHE), Fig. 7.16, are mostly fin-and-tube or plate-type designs where
advances in ceramic materials are providing novel designs for high temperature applications [369].
According to Li et al. [369] the ceramic material is defined as a glazed/unglazed body of crystalline
or partly crystalline structure, or of glass. Here the structure is produced from essentially inorganic,
nonmetallic substances and is either formed from molten mass which solidifies on cooling [399].

Figure 7.16. Industrial ceramic heat exchanger [400].

Currently, CHEs are primarily constructed by replacing tube and fin materials with ceramic.
Manufacturing procedures include forming primary components from raw materials with
subsequent machining, joining, bonding, and assembling, in either a monolithic and non-
monolithic approach. Because individual components are bonded together permanently without an
internal joint, there are no sealing problems in a monolithic assembly, but stress concentrations
can arise under extreme operating conditions. In a non-monolithic assembly, individual
components can be disassembled and repaired easily, however sealing issues can arise due to the
brittleness of ceramics and challenges of forming a resilient metal–ceramic or ceramic–ceramic
joint due coefficient of thermal expansion (CTE) differences [401]. Consequently, advanced
fabrication techniques like laser supported brazing [402] are often employed to improve joint
integrity. Primary advantages for using ceramic materials over traditional metallic materials in

154
CHE construction are their extremely high temperature stability, low material cost and excellent
corrosion resistance. CHEs can withstand operating temperatures that potentially can far exceed
those of conventional metallic alloys, Silicon carbide (SiC) 1400 ᵒC, Silicon nitride (Si3N4) 1900
°C, Alumina 1500–1700 ᵒC, and Aluminium nitride (AlN) 1300 ᵒC [401]. At elevated operating
temperatures, CHEs possess high fouling resistance and resistance to corrosion and chemical
erosion.
Among the ceramics materials for CHEs, SiC ceramics that have been extensively
investigated [403], which has been treated as a promising material for high-temperature heat
exchanger applications, primarily because of its excellent thermal stability and corrosion resistance
in severe environments, as well as sufficient thermal conductivity at elevated temperatures [404].
Steen [405] investigated the short and long term mechanical properties of a sintered silicon carbide
heat exchanger within advanced gas turbine coal-fired power plants. Amirthan et al. [406] prepared
four different types of Si/SiC ceramic composites by a liquid-silicon infiltration technique and
measured their thermal conductivities at different temperatures by a laser flash thermal
conductivity method. The investigators found that the presence of free carbon and voids notably
affect the thermal conductivity of these materials. Investigations of ceramic CHEs involved in
liquid-to-liquid heat exchangers [407], liquid-to-gas heat exchangers and gas-to-gas heat
exchangers [408] could be leveraged for use in heat sinks and volumetric solar receivers. Schulte-
Fischedick et al. [409] proposed a sintered SiC ceramic PFHE coated with an environmental barrier
coating based on the “Offset Strip Fin” design, where the investigators studied it for use in a high
temperature heat exchanger for an externally-fired combined cycle (EFCC), as well as for other
applications that required extreme temperature operation up to 1250 °C. The thermal design was
conducted by using empirical correlations drawn from literature via the LMTD-method, while the
stress distribution of the selected design during operation was investigated by means of finite
element method (FEM). Islamoglu [408] analyzed the temperature distribution for steady-state
heat transfer, as well as the thermal stresses, where were induced by temperature difference in a
silicon carbide (SiC) ceramic tube. The employment effects of annular fins with different profiles
(rectangular and triangular) [410] and the axial non-uniform convective heat transfer coefficient
[408] on the temperature and thermal stresses have also been investigated. Fend et al. [411]
measured thermal conductivity, convective heat transfer coefficients and efficiency of selected
materials made from various oxide/non-oxide ceramics, ceramic foams and metal structures, where
they also experimentally investigated how the properties of the porous materials affect flow
stability. It was concluded that introduction of materials based on ceramic foams and ceramic
fabrics in solar tower technology would offer the potential for more effective and reliable operation
compared to current open volumetric receiver technology. They presented test results for two novel
porous materials, for an application as a volumetric receiver using a double-layer silicon carbide
foam and a screen-printed porous silicon carbide material. Both ceramic technologies are future
options with properties required for use in a high-performance solar receiver [412].

7.1.9 Fluidized-Bed Heat Exchangers

For fluidized-bed heat exchangers one characteristic type consists of tube bundles immersed in a
bed of finely divided solid material, as shown in Fig. 7.17. If the upward fluid velocity on the bed
side is low, the solid particles will remain fixed within the bed and the fluid will flow through the

155
interstices of the bed. If the upward fluid velocity is high, the solid particles will be carried away
with the fluid.

Figure 7.17. Fluidized-bed heat exchanger.

At high fluid velocities, the upward drag force is slightly higher than the weight of the bed particles
which will cause the solid particles to float with an increase in bed volume causing the bed to
behave as a liquid [413]. This is referred to as a fluidized condition where the pressure drop through
the bed remains almost constant and independent of the flow rate. Under this condition a strong
mixing of the solid particles occurs, especially for small particle diameters that are less than 1 mm,
which results in a homogeneous temperature distribution of both gas and solid media [359]. Very
high heat transfer coefficients are achieved on the fluidized-side compared to particle-free or
dilute-phase particle gas flows. Chemical reaction can also be facilitated on the fluidized side
depending on the level of oxygen and activation energy of the particles [414]. Since the initial
temperature difference, between the hot fluid and fluidized bed, Th,0-Tf,0 is reduced due to
fluidization, the exchanger effectiveness is therefore lower [415].
For CSP combustion applications, input media with high reactivity have greater
combustion or reaction efficiency, where previous studies have demonstrated the amount of
combustible losses in fly ash is inversely proportional to the heating value of the fuel [417]. Typical
combustion temperatures for fluidized beds falls within the range of 800 – 900 ᵒC [417]. Although
this temperature range minimizes the formation of sulfur and nitrogen oxides, higher temperatures
will ensure reduced combustible losses in escaping unburned fuel particles with incomplete
combustion. In fluidized bed combustors, the residence time of carbon particles is affected by the
type of feed, particle size and fluidizing velocity. The residence time becomes shorter for particles
with significantly lower velocities, favoring a high percentage of unburned carbon [417].
Principally, when an upward flowing gas moving with a relative velocity is passed through
a static bed of particles, they will be partially suspended once the gas velocity reaches a certain
minimum fluidization velocity [418], therefore the minimum fluidization velocity is that required
to move the static bed. The quality of fluidization obtained varies from one bed media to another,
where intrinsic properties including particle density, particle size and surface characteristics affect
the outcome of fluidization [419]. The operating velocity of the fluidized bed must be maintained

156
between the minimum fluidization velocity and the terminal velocity, which is the maximum gas
velocity beyond which particles will be entrained.

7.2 Recuperators

In most heat exchangers, heat transfer between fluids takes place through a separating wall, or into
and out of a wall in a transient manner, where the fluids are separated by a heat transfer surface.
However, in some design configurations direct-transfer type exchangers are required to enhance
heat transfer between fluids where one fluid recovers (recuperates) heat from a second fluid. Some
examples of recuperators can be seen in Fig. 7.18, where direct transfer designs can also include
tubular, plate-type, and extended surface exchangers. Recuperators can be added between pump
and expander outlets to preheat a liquid before the evaporator, which increases cycle efficiency.

a. b.
Figure 7.18. Generalized recuperator heat transfer modes.

7.3 Regenerators

In contrast to recuperators, regenerator heat exchangers are storage-type exchangers, which

facilitate intermittent heat exchange between hot and cold fluids, via thermal energy storage and
release through an exchanger surface or matrix. As shown in Fig. 7.19, there are several
characteristic variations that transfer heat indirectly, where these devices usually have fluid
interaction from one fluid stream to the other due to pressure differences or matrix rotation/valve
switching. In general, a stationary heat exchanger is usually referred to as a recuperator, and a
dynamic or rotating heat exchanger as a regenerator [359].

157
 b.

a.

c.
Figure 7.19. Regenerator types: a. rotary, b. fixed-matrix and c. Rothemuhle [169].

According to Kreith [169] in order to facilitate continuous regenerator operation, the internal
matrix must be moved periodically within the fixed streams of gases, as in a rotary regenerator, or
the gas flows must be diverted through valves to and from the fixed matrices. This can be
accomplished with a fixed-matrix regenerator, also known as a periodic-flow regenerator or
reversible heat accumulator. Rotary regenerators are advantageous for highly compact forms,
where the cost of the regenerator surface per unit of heat transfer area is usually substantially lower
than that for the equivalent recuperator. A major disadvantage of a regenerator is an unavoidable
carryover of a small fraction of the fluid trapped in the passage to the other fluid stream just after
periodic flow switching [169]. Since fluid contamination (small mixing) is prohibited with liquids,
the regenerators are used exclusively for gas-to-gas heat or energy recovery applications. Cross
contamination can also be minimized significantly by providing a purge section in the disk and
using double-labyrinth seals [169]. Rotary regenerators have been designed for a surface area
density of up to approximately 6600 m2/m3, with an exchanger effectivenesses exceeding 85% for
a number of applications. These devices can additionally employ thinner stock material, resulting
in minimal volumes of material necessary for a given effectiveness and pressure drop. Metal rotary
regenerators have been designed for continuous inlet temperatures of up to approximately 790 °C,

158
and pressures differences up to 400 kPa between hot and cold gases, where ceramic matrices can
potentially achieve even higher temperature applications [169]. Finally, another type of
regenerator is a Rothemuhle regenerator which is used commonly as an air preheater in some
power-generating plants, which has a fixed matrix, where the gases pass through rotating hoods to
the matrix [169].
In storage heat exchangers both fluids flow alternatively through the same flow passages,
where heat transfer is intermittent. The heat transfer flow passages generally have a cellular
structure (matrix), or is a permeable solid material (packed bed). When hot gas flows over the heat
transfer surface through flow passages, thermal energy is stored in the matrix wall, where the gas
is cooled during the matrix heating period. As cold gas flows through the same passages thereafter,
it absorbs thermal energy from the matrix wall. This storage type heat exchanger is also referred
to as a regenerator. The actual time hot gas takes to flow through a cold regenerator matrix is called
the hot period, and the time that cold gas flows through the hot regenerator matrix is called the
cold period. There is a small fraction of carryover of the fluid trapped in the passage to the other
fluid stream just after switching of the fluids, referred to as carryover leakage. In addition, if the
hot and cold fluids are at different pressures, pressure leakage from the high-pressure fluid to the
low-pressure fluid past the radial, peripheral, and axial seals, or across the valves. Over time leaks
can be unavoidable where regenerators are used exclusively in gas-to-gas heat transfer sensible
applications. Finally, in some applications, regenerators may transfer moisture from humid air to
dry air up to approximately 5% [169].

159
8. CONCLUDING REMARKS
The novelty of CSP is that it captures and concentrates the sun’s incident irradiant energy to
provide heat required to generate electricity, where CSP plants can be equipped with heat storage
to generate electricity even when the sky is cloudy or dark. This provides tremendous potential for
impacting the base load of an electric grid by significantly increasing the capacity factor compared
with solar photovoltaics by providing dispatchable electricity, which can facilitate both grid
integration and economic competitiveness. Heat transfer in solar thermal systems are guided by
five basic principles: heat gain; thermal-fluid transport; heat storage; heat losses; and heat
insulation. CSP maximum efficiency can be enhanced and shifted towards higher temperatures
when solar irradiation is further concentrated by CSP collectors, where high-absorption receivers
are able to transfer large thermal loads to a heat transfer fluid for power block electricity
generation. CSP technologies today benefit from new field concentrator layouts and optics, as well
as advances in heat transfer to achieve even higher concentration ratios, as well as thermal storage
technologies, while other components of CSP plants can benefit from thermal-fluid advances in
heat exchangers and power block system technologies. Although CSP performance increases with
higher temperatures, thermal losses also increase however especially at temperatures above 600
°C where radiation effects can become dominant for most surfaces. The utility of fundamental heat
transfer phenomena at varying temperature/pressure scales has contributed to enhanced HTF
materials that span solids, liquids and gases, which can achieve higher material temperatures while
remaining stable. Novel supercritical gases have also been explored that can attain properties close
to those of liquids for enhanced energy transfer, storage and power block system interactions.
Novel development of HTFs has contributed to enhancement of their ability to also store and
transfer heat through means of novel heat exchange equipment that also allow for higher thermal
absorption to achieve even higher thermal efficiencies. However, system component equipment
and HTF limitations still exist, such as losses, which require further heat transfer study. However,
current advancements have contributed to the ability of transferring and retaining heat within
innovative thermal cycles. Such cycles like the supercritical Brayton cycle, take advantage of
components such as regenerators, recuperators and cascaded designs, as well as novel materials
and coatings. These and other advancements demonstrate the potential for achieving even higher
temperatures and pressures to facilitate further LCOE reductions for CSP competitiveness, to
enable cost-effective energy sufficiency in a carbon-neutral future.

160
9. REFERENCES
[1] Krothapalli, A. and Greska, B. "Concentrated solar thermal power." USA: Florida State
University, 2011.
[2] IRENA Report, Concentrating Solar Power, Technology Brief E10, IEA-ETSAP, 2013.
[3] Fichtner, “Technology Assessment of CSP Technologies for a Site Specific Project in South
Africa Final Report,” The World Bank and ESMAP, Washington D.C., 2010.
[4] Henry, A., & Prasher, R., The prospect of high temperature solid state energy conversion to
reduce the cost of concentrated solar power. Energy & Environmental Science, 7, No. 6, pp.
1819-1828, 2014.
[5] Hinkley, J., Curtin, B., Hayward, J., Wonhas, A., Boyd, R., Grima, C., ... & Mikhail, A. (2011).
Concentrating solar power–drivers and opportunities for cost-competitive electricity. Clayton
South: CSIRO.
[6] Hoffschmidt, B., “Receivers for solar tower systems”, SFERA, Font Romeu, France, 2014
[7] Mehos, M., Turchi, C., Jorgenson, J., Denholm, P., Ho, C., & Armijo, K.M., “On the Path to
SunShot: Advancing Concentrating Solar Power Technology, Performance, and
Dispatchability”, No. NREL/TP-5500-65688; SAND2016-2237 R, Golden, CO, USA, 2016
[8] Molenbroek, E., “Concentrating Solar Power Status and Potential”, ECOFYS
Report, 2008.
[9] Duffie, J.A., Beckman, W.A., “Solar engineering of thermal processes”, 4th Ed., Wiley & Sons,
New Jersey, USA, 2013.
[10] Armijo, K. M., Harrison, R. K., King, B. H., & Martin, J. B. (2014, September). Spectral
derates phenomena of atmospheric components on multi-junction CPV technologies. In AIP
Conference Proceedings (Vol. 1616, pp. 264-271).
[11] Iqbal, M., An Introduction to Solar Radiation, Academic, Toronto, 1983.
[12] Incropera, F. P., & Dewitt, D. P., “Fundamentals of Mass and Heat Transfer”, Wiley & Sons,
New Jersey, USA, 2002
[13] Ibarra-Castanedo, C., Sfarra, S., Genest, M., & Maldague, X., “Infrared vision: Visual
inspection beyond the visible spectrum”, In: “Integrated imaging and vision techniques for
industrial inspection”, pp. 41-58. Springer London, 2015.
[14] Duffie, J. A., and Beckman, W. A., “Solar engineering of thermal processes” 4th Edition,
Wiley & Sons, New York, USA, 2013.
[15] Tada, H., “Gourmet Engineering Lecture Notes”, Tufts School of Engineering, 2002.
[16] Chirkin, V. S., “Thermal conductivity of industrial materials”, Mashgiz, Moscow, 1962.
[17] Andraka, C. E. (2014). Dish Stirling advanced latent storage feasibility. Energy Procedia, 49,
684-693.
[18] Hashmi, A., Tahir, F. R. A. A. Z., and Hameed, U. “Empirical Nusselt number correlation for
single phase flow through a plate heat exchanger”, In” Proceedings of the 9th WSEAS
International Conference on Heat and Mass Transfer, Cambridge, MA, USA, pp. 25-27,
2012.
[19] J. Fernandez-Seara, J., Sieres, U.J., Campo, A., “A general review of the Wilson plot method
and its modifications to determine convection coefficients in heat exchange devices”, Applied
Thermal Engineering, 27, pp. 2745–2757, 2007.
[20] Edwards, M.F., Changal Vaie, A.A., and Parrott, D.L., “Heat transfer and pressure drop
characteristics of a plate heat exchanger using non-Newtonian liquids”, The Chemical
Engineer, 285, pp. 286–288, 293, 1974.

161
[21] Muley A. and Manglik, R.M., “Experimental study of turbulent flow heat transfer and
pressure drop in a plate heat exchanger with chevron plates”, J. of Heat Transfer, 121, no. 1,
pp. 110–117, 1999.
[22] Gnielinski, V. “New equations for heat and mass-transfer in turbulent pipe and channel flow”,
International chemical engineering, 16, no.2, 359-368, 1976.
[23] Gourdon, M., Innings, F., Jongsma, A., & Vamling, L., “Qualitative investigation of the flow
behaviour during falling film evaporation of a dairy product”, Experimental Thermal and
Fluid Science, 60, pp. 9-19, 2015.
[24] Zhukauskas, A., “Heat transfer from tubes in cross flow”, In: J.P. Hartnett and T.F. Irvine Jr.,
Eds., Advances in Heat Transfer, 8, Academic Press, New York, USA, 2972.
[25] Carey, V.P., “Liquid-vapor phase-change phenomena”, 2nd Ed., Taylor & Francis, New York,
USA, 2008.
[26] Ramé-Hart, “Contact angle goniometers and tensionometers”, Ramé-Hart Instrument Co.,
2015: http://www.ramehart.com/contactangle.htm
[27]Eistron, “Contact angle meter”, Weiston Co. LTD., 2014:
http://www.weistron.com/products/contact-angle-meter/
[28] Pfleger, N., Bauer, T., Martin, C., Eck, M., and Wörner, A. “Thermal energy storage–
overview and specific insight into nitrate salts for sensible and latent heat storage”, Beilstein
journal of nanotechnology, 6, No. 1, pp. 1487-1497, 2015.
[29] Rohsenow, W.M., “A method of correlating heat transfer data for surface boiling of liquids”,
Trans. ASME, 84, pp. 969–975, 1962.
[30] Faghri, A., Zhang, Y., & Howell, J. R., “Advanced heat and mass transfer”, Global Digital
Press, Columbia, MO, USA, 2010.
[31] Dwyer, O. E., “Boiling liquid-metal heat transfer”, American Nuclear Society, Hinsdale, Ill.,
1976.
[32] Subbotin, V.I., Ushakov, P.A., Kirillov, P.L., Ibragimov, M.K., Ivanovski, M.N, Nomophilov,
E.V., Ovecnkin, D.M., Sorokin, D.N. and Sorokin, V.P., “Heat removal from the reactor fuel
elements cooled by liquid metals”, In: Proc. UN Intern. Conf. Peaceful Uses of Atomic energy,
Geneva, Sept. 3rd 1964, 8, pp. 192, United Nations, NY, USA, 1965.
[33] Edwards, J.A. and Hoffman, H.W., “Superheat with boiling alkali metals”, In: Proc. Conf.
Application of high-temperature instrumentation to liquid-metal experiments, Argonne
national Laboratory, ANL-7100, 1065.
[34] Chen, J.C., “Incipient boiling superhets in liquid metals”, Trans. ASME Series C, J. of Heat
Transfer, 90, 1968.
[35] Logan, A.D., Baroczy, C.J., Landoni, J.A. and Morewitz, H.A., “Effects of velocity, oxide
level, and flow transients on boiling initiation in sodium”, Proc. ASME Symp. Liquid-Metal
Heat Transfer and Fluid Dynamics, pp. 116, 1970.
[36] Collier, J.G. and Thome, J.R., “Convective boiling and condensation”, 3rd Ed., Oxford
University Press, Oxford, U.K., 1994.
[37] Wright, R.M., “Downflow forced convection boiling of water in uniformly heated tubes”,
Univ. of California, Report No. UCRL-9744, 1961.
[38] Gungor, K.E. and Winterton, R.H.S., “A general correlation for flow boiling in tubes and
annuli”, Int. J. Heat & Mass Transfer, 29, pp. 351-358, 1986.
[39] Bjorge, R.W., Hall, G.R. and Rohsenow, W.M., “Correlation of forced convection boiling
heat transfer data”, Int. J. Heat & Mass Transfer, 25, pp. 753-757, 1982.

162
[40] Kandlikar, S.G., “A general correlation for saturated two-phase flow boiling heat transfer
inside horizontal and vertical tubes”, J. of Heat Transfer, 112, pp. 219-228, 1990.
[41] Kandlikar, S.G., “Development of a flow boiling map for subcooled and saturated flow
boiling of different fluids in circular tubes”, J. of Heat Transfer, 113, pp. 190-200, 1991.
[42] Gunther, F.C., “Photographic study of surface-boiling heat transfer to water with forced
convection”, Trans. ASME, 73, pp. 105-107, 1951.
[43] Tong, L.S., “Boundary layer analysis of the flow boiling crisis”, Int. J. of Heat and Mass
Transfer, 11, pp. 1208-1211, 1968.
[44] Tong, L.S., “A pheonomenological study of critical heat flux, ASME paper 75-HT-68,
National Heat Transfer Conference, Sn Francisco, CA, USA, 1975.
[45] Thompson, B. and Macbeth, R.V., “Boiling water heat transfer – Burnout in uniformly heated
round tubes: A compilation of world data with accurate correlations”, Report AEEW-R-356,
U.K. Atomic Energy Authority, Winfrith, England, 1964.
[46] Biasi, L., Clerici, G.C., Gariloben, S., Sala, R. and Tozzi, A., “Studies on burnout. Part 3”,
Energia Nucleare, 14, No. 9, pp. 530-536, 1967.
[47] Katto, Y. and Ohno, H., “An improved version of the generalized correlation of critical heat
flux for the forced convective boiling in uniformly heated vertical tubes”, Int. J. of Heat and
Mass Transfer, 27, pp. 1641-1648, 1984.
[48]Stine, W. and Geyer, M., “Power from the sun”, Ch. 9, 2001,
http://www.powerfromthesun.net/Book/chapter09/chapter09.html
[49] Duffie, J. A. and Lof, G.O.G., ‘‘Focusing Solar Collectors for Power Generation’’, Paper 207
III.7/5, presented at the World Power Conference, Melbourne, Australia, 1962.
[50] Grange, B., “Solar Receiver modeling central receiver”, SFERAII Presentation, Odeillo,
France, 2014.
[51] Treadwell, G. W. “Design Considerations for Parabolic Cylindrical Solar Collectors”,
SAND76-0082, Sandia National Laboratories, Albuquerque, 1976.
[52] Ratzel, A C. “Evaluation of the Evacuated Solar Annular Receivers Employed at the
Midtemperature Solar Systems Test Facility”, SAND78-0983, Sandia National Laboratories,
Albuquerque, 1979.
[53] Ratzel, A. C. “Receiver Assembly Design Studies for 2-m 90 Degree Parabolic Cylindrical
Solar Collectors”, SAND79-1026, Sandia National Laboratories, Albuquerque, 1979.
[54] Ratzel, A.C., and C.E. Simpson, “Heat Loss Reduction Techniques for Annular Receiver
Designs”, SAND 78-1769, Sandia National Laboratories, Albuquerque, 1979.
[55] Kalogirou, S.A., “Solar thermal collectors and applications”, Prog. Energy & Comb. Sci., 30,
pp. 231-295, 2004.
[56]Stine, W. and Geyer, M., “Power from the sun”, Ch. 5, 2001,
http://www.powerfromthesun.net/Book/chapter05/chapter05.html
[57] Bejan, A., Kearney, D.W. and Kreith, F., “Second law analysis and synthesis of solar collector
systems”, J. of Solar Energy Engng., 103, pp. 23-28, 1981.
[58] Bejan, A., “Entropy generation minimization”, 2nd Ed., CRC Press, Boca Raton, FL, USA,
1995.
[59] Petela, R., “Exergy of heat radiation”, ASME J Heat Transfer, 68, pp. 187, 1964.
[60] McDonald, G. E. “Spectral reflectance properties of black chrome for use as a solar selective
coating”, Solar Energy, 17, No. 2, pp.119-122, 1975.
[61] Hollands, K. G. T., ‘‘Directional selectivity, emittance and absorptance properties of Vee
Corrugated Specular Surfaces’’, Solar Energy, 7, pp. 108, 1963.

163
[62] Trombe, F., Foex, M. and LePhat Vinh, M., ‘‘Research on Selective Surfaces for Air
Conditioning Dwellings.’’, Proc. UN Conf. New Sources of Energy, 4, pp. 625-638, 1964.
[63] V. E. Dudley, G. J. Kolb, M. Sloan, and D. Kearney, “Test Results: SEGS LS-2 Solar
Collector”, SAND94-1884, Sandia National Laboratories, Albuquerque, NM, USA, 1994.
[64] Ho, C. K., & Iverson, B. D., “Review of high-temperature central receiver designs for
concentrating solar power”, Renewable and Sustainable Energy Reviews, 29, pp. 835-846,
2014.
[65] Prakash, M., Kedare, S. B., & Nayak, J. K., “Investigations on heat losses from a solar cavity
receiver”, Solar Energy, 83, No. 2, pp.157-170, 2009.
[66] Kolb, G. J., Ho, C. K., Mancini, T. R., & Gary, J. A. “Power tower technology roadmap and
cost reduction plan”, SAND2011-2419, Sandia National Laboratories, Albuquerque, NM,
2011.
[67] Becker, M. and Vant-Hull, L. L., “Thermal receivers, Solar Power Plants”, Springer, Berlin,
Germany, 1991.
[68] Kribus, A., Doron, P., Rubin, R., Karni, J., Duchan, S. and Taragan, E., “A multistage solar
receiver: the route to high temperature”, Solar Energy, 67, No. 1, pp.3–11, 1999.
[69] Kugath, D. A. Drenker, G. and Koenig, A. A. “Design and development of a paraboloidal
dish solar collector for intermediate temperature service,” Proc. of Int. Solar Energy Soc.,
Silver Jubilee Congress, 1, pp. 449–453, Atlanta, Ga, USA, 1979.
[70] Kedare, S. B., Nayak, J. K. and Paranjape, A. D., “Development, installation and evaluation
of large scale concentrating solar collector for medium temperature industrial thermal
applications,” Final Report of R&D Project No. 15/9/2002-ST, Ministry of New and
Renewable Energy, Government of India, 2006.
[71] Lovegrove, A. Zawadski, and J. Coventy, “Taking the anu big dish to commercialization,”
Proc. of Solar 2006: Clean Energy? -Can Do!, (ANZSES ’06), Canberra, Australia, 2006.
[72] F. Nepveu, A. Ferriere, and F. Bataille, “Thermal model of a dish/stirling systems,” Solar
Energy, 83, No. 1, pp. 81–89, 2009.
[73] A. Steinfeld and R. Palumbo, “Solar thermochemical process technology,” Encyclopedia of
Physical Science & Technology, 15, pp. 237–256, 2001.
[74] Harris, J. A., & Lenz, T. G., “Thermal performance of solar concentrator/cavity receiver
systems”, Solar Energy, 34, No.2, pp. 135-142, 1985.
[75] Dudley, V. E., and Workhoven, R. M., “Performance Testing of the Suntec- SLATS Solar
Collector,” SAND78-0623, Sandia National Laboratories, Albuquerque, NM, USA, 1978.
[76] Ma, R. Y. “Wind effects on convective heat loss from a cavity receiver for a parabolic
concentrating solar collector,” SAND92-7293, Sandia National Laboratories, Albuquerque,
NM, USA, 1993.
[77] Prakash, M., “Numerical study of natural convection heat loss from cylindrical solar
cavity receivers,” ISRN Renewable Energy, 2014.
[78] Taumoefolau, T., Paitoonsurikarn, S., Hughes, G. and Lovegrove, K.,“Experimental
investigation of natural convection heat loss from a model solar concentrator cavity
receiver,” J. of Solar Energy Eng., 126, No. 2, pp. 801–807, 2004.
[79] Sendhil Kumar, N. and Reddy, K.S., “Numerical investigation of natural convection heat
loss in modified cavity receiver for fuzzy focal solar dish concentrator,” Solar Energy, 81,
no. 7, pp. 846–855, 2007.
[80] Stine, W.B., and McDonald, C.G., “Cavity receiver convective heat loss,” Proc. of Int.
Solar Energy Soc. Solar World Cong., 2, pp. 1318–1322, Kobe, Japan, 1989.

164
[81] Sendhil Kumar, N. and Reddy, K. S., “Comparison of receivers for solar dish collector
system,” Energy Conversion and Management, 49, No. 4, pp. 812–819, 2008.
[82] Kistler, B. L. “A user’s manual for DELSOL3: a computer code for calculating the optical
performance and optimal system design for solar thermal central receiver plants,” SAND86-
8018, Sandia National Laboratories, Albuquerque, NM, USA, 1986.
[83] Koenig, A.A., Marvin, M., “Convection heat loss sensitivity in open cavity solar receivers,”
Final Report, DOE Contract No. EG77-C-04-3985, D.O.E., Oak Ridge, Tennessee, USA,
1981.
[84] Siebers, D.L. and Kraabel, J.S., “Estimating convective energy losses from solar central
receivers,” SAND84-8717, Sandia National Laboratories, Albuquerque, NM, USA, 1984.
[85] Kraabel, J.S., “An experimental investigation of natural convection from a side-facing cubical
cavity,” ASME-JSME Thermal Eng. Joint Conf. Proc., 1, pp. 299-306, Honolulu, Hawaii,
USA, 2983.
[86] LeQuere, P., Penot, F. and Mirenayat, M., “Experimental study of heat loss through natural
convection from an isothermal cubic open cavity,” SAND81-8014, Sandia National
Laboratories, Albuquerque, NM, USA, 1981.
[87] Clausing, A.M., “An analysis of convective losses from cavity solar central receviers,” Solar
Energy, 27, No. 4, pp. 295-300, 1981.
[88] McMordie, R.K., “Convection heat loss from a cavity receiver, “ASME J. of Solar Energy
Eng., 106, pp. 98-100, 1984.
[89] Faust, K.M., Plate, E.J. and Kuezera, M., “Experimental Investigation of the convective losses
from the cavity receiver of the project GAST,” SAND81-8014, Sandia National Laboratories,
Albuquerque, NM, USA, 1981.
[90] Paitoonsurikarn, S., Lovegrove, K., Hughes, G. and Pye, J., “Numerical investigation of
natural convection loss from cavity receivers in solar dish applications,” J. of Solar Energy
Eng., 133, No. 2, 2011.
[91] Prakash, M., Kedare, S. B., & Nayak, J. K., “Determination of stagnation and convective
zones in a solar cavity receiver,” Int. J. of Therm. Sci., 49, No. 6, pp. 680-691, 2010.
[92] Flamant, G., Gauthier, D., Benoit, H., Sans, J. L., Garcia, R., Boissière, B. and Hemati, M.
“Dense suspension of solid particles as a new heat transfer fluid for concentrated solar thermal
plants: on-sun proof of concept”, Chemical Engineering Science, 102, pp. 567-576, 2013.
[93] Röger, M., Amsbeck, L., Gobereit, B. and Buck, R., “Face-down solid particle receiver using
recirculation,” J. of Solar Energy Eng., 133, No. 3, 2011.
[94] Wu, W., Gobereit, B., Singer, C., Amsbeck, L. and Pitz-Paal, R., “Direct absorption receivers
for high temperatures,” 2011.
[95] Ho, C. K., Christian, J. M., Yellowhair, J., Siegel, N., Jeter, S., Golob, M., and Al-Ansary,
H., “On-sun testing of an advanced falling particle receiver system,” SOLARPACES 2015:
International Conference on Concentrating Solar Power and Chemical Energy Systems,
1734, No. 1, 2016.
[96] Chen, H., Chen, Y., Hsieh, H. T., and Siegel, N., “Computational fluid dynamics modeling
of gas-particle flow within a solid-particle solar receiver,” Journal of solar energy
engineering, 129, No. 2, pp. 160-170, 2007.
[97] Penner, A., “High temperature falling particle receiver reaches new limits,” Sandia National
Laboratories: Energy & Climate, Accessed: August, 2015, http://energy.sandia.gov/high-
temperature-falling-particle-receiver-breaks-reaches-new-limits/.

165
[98]. Jones, P.D., Radiation and convection heat trasnfer in particle-laden fluid flow. Dissertation
Doctoral Thesis, Rice University, 1991.
[99]. R. Yamazaki, N.U.a.G.J., “Mechanism of incipient fluidization and fluidized bed at elevated
temperature”, J. of Chemical Engineering of Japan19: p. 251-257, 1986.
[100]. C.E. Davies, S.G.B.D.a.R.B.F., “Fluidization VI. An investigation of thermal agglomeration
in fluidized beds”, Engineering Foundation, New York, USA, 1989.
[101]. Shimizu, T.T.a.A., “Advances in Multiphase Flow. Prediction of turbulence modulation by
using k-e model for gas-solid flow,” Elsevier Science, Amsterdam, NL,1995.
[102]. Basu, P.K.N., “An investigation into heat trasnfer in CFB's,” J. of Heat and Mass Trans.,
30, p. 2399, 1987.
[103]. C. Heered, A.P.V.a.D.W.V., “Mechanism of heat trasnfer in FBs,” Industrial and Chemistry
Research, 45, No. 6, p. 1237, 1953.
[104]. Kolar, M.M.a.A.K., “Emulsion layer model for wall heat transfer in CFB,”AlChe Journal,
37, p. 1139, 1991.
[105]. Mickley, D.F.F., “Mechanism of heat trasnfer in FBs,” AlCheE Journal, 1, p. 374, 1960.
[106]. Basu, D.S.a.P., “A model for heat transfer in CFBs,” Int. J. of Heat and Mass Transfer, 29:
p. 487, 1986.
[107]. Williams, J.S.B.a.J.R., “The mechanism of heat transfer in gas-FBs,” Trans. of Inst. of
Chem. Eng., 41, p. 217, 1963.
[108]. Martin, H., “Heat transfer between gas fluidized beds of solid particles and the surfaces of
immersed heat exchanger elements Part II,” Chem. Eng. & Proc.: Proc. Intensification, 18,
No. 4, p. 199-223, 1984.
[109]. Drake, E.R.G.E.a.R.M., “Analysis of Heat and Mass Transfer,” McGraw-Hill, New York:
1972.
[110]. Miller, C.Y.W.a.E.N., “Heat transfer in solid-gas transport lines,” J. of Ind. and Eng. Chem.,
53, p. 51, 1961.
[111] Ma, Z., Glatzmaier, G., & Mehos, M., “Fluidized bed technology for concentrating solar
power with thermal energy storage,” Journal of Solar Energy Engineering, 136, No. 3, 2014.
[112] Zhou, Z., Hou, Q. and Yu, A., “Particle scale simulation of heat transfer in fluid bed
reactors,” INTECH Open Access Publisher, Ch. 15, 2011.
[113] Loth, E., Tryggvason, Tsuji, Y., Elghobashi, S.E., Crowe, C.T., Berlemont, A., Reeks, M.,
Simonin, O., Frank, Th., Onishi, Y., van Wachem, B., “Multiphase Flow Handbook:
Modeling”, Ch. 13, Taylor & Francis, Boca Raton, FL, 2006.
[114] Anderson, T.B. and Jackson, R., “Fluid mechanical description of fluidized bed sequations
of motion,” Ind. Eng. Chem. Fund., 6, pp. 527-539, 1967.
[115] Ishii, M., “Thermo-fluid dynamic theory of two-phase flow,” Eyrolles, Paris, FR, 1975.
[116] Gidaspow, D., “Multiphase Flow and Fluidization,” Academic Press, San Diego, USA,
1994.
[117] Tsuji, Y., Tanaka, T. & Ishida, T., “Lagrangian numerical simulation of plug flow of
cohesionless particles in a horizontal pipe,”Powder Technol., 71, pp. 239-250, 1992.
[118] Xu, B.H. & Yu, A.B., “Numerical simulation of the gas-solid flow in a fluidized bed by
combining discrete particle method with computational fluid dynamics,” Chem. Eng. Sci., 52,
1997.
[119] Crocker, A. and Miller, F., “Fluid flow and radiation modeling of a cylindrical small particle
solar receiver,” SolarPACES 2011, Granada, Spain, 2011.

166
[120] Ho, C.K., “Sandia advanced concepts,” Technical Report 505, Sandia National Laboratories,
Washington D.C., 2010.
[121] Zhou, Z.Y., Yu, A.B. and Zulli, P., “Particle scale study of heat transfer in packed and
bubbling fluidized beds,” AIChE J., 55, pp.868-884, 2009.
[122] Zhou, Z.Y., Yu, A.B. and Zulli, P., “A new computational method for studying heat transfer
in fluid bed reactors,” Powder Technol., 197, pp. 102-110, 2010.
[123] Zhu, H.P., Zhou, Z.Y., Yang, R.Y. and Yu, A.B., “Discrete particle simulation of particulate
systems: Theoretical developments,” Chem. Eng. Sci., 62, 2007.
[124] Ergun, S., “Fluid flow through packed columns,” Chem. Eng. Prog., 48, pp. 89-94, 1952.
[125] Di Felice, R., “The voidage function for fluid-particle interaction systems,” Int. J.
Multiphase Flow, 20, pp. 153-159, 1994.
[126] Wakao, N. and Kaguei, S., “Heat and Mass Transfer in Packed Beds,” Gordon and Breach
Science Publishers, New York, USA, 1982.
[127] Kunii, D. and Levenspiel, O., “Fluidization Engineering,” Butterworth-Heinemann, 978-0-
409-90233-4, Boston, USA, 1991.
[128] Holman, J.P., “Heat Transfer,” McGraw-Hill Company, New York, USA, 1981.
[129] Yagi, S. and Kunii, D., “Studies on effective thermal conductivities in packed beds,” AIChE
J., 3, pp. 373-381, 1957.
[130] Cheng, G.J., Yu, A.B. and Zulli, P., “Evaluation of effective thermal conductivity from the
structure of a packed bed,” Chem. Eng. Sci., 54, 1999.
[131] Sun, J. & Chen, M.M., “A theoretical analysis of heat transfer due to particle impact,” Int.
J. Heat Mass Transfer, 31, pp. 969-975, 1998.
[132] Parmar, M.S. and Hayhurst, A.N., “The heat transfer coefficient for a freely moving sphere
in a bubbling fluidized bed,” Chem. Eng. Sci., 57, 2002.
[133] Collier, A.P., Hayhurst, A.N., Richardson, J.L. and Scott, S.A., “The heat transfer coefficient
between a particle and a bed (packed or fluidised) of much larger particles,” Chem. Eng. Sci.,
59, 2004.
[134] Scott, S.A., Davidson, J.F., Dennis, J.S. and Hayhurst, A.N., “Heat transfer to a single sphere
immersed in beds of particles supplied by gas at rates above and below minimum
fluidization,” Ind. Eng. Chem. Res., 43, pp. 5632-5644, 2004.
[135] Armijo, K. M., Ho, C., Anderson, R., Christian, J., Babiniec, S., and Ortega, J., “Magnetic
field flow phenomena in a falling particle receiver,” SOLARPACES 2015, 1734, No. 1, AIP
Publishing, 2006.
[136] E. Gedik, H. Kurt and Z. Recebli, “CFD simulation of magnetohydrodynamic flow of a
liquid-metal Galinstan fluid in circular pipes,” FDMP, 9, pp. 23-33, 2013.
[137] Nakatsuka, K., Jeyadevan, B., Neveu S., and Koganezawa, H., “The magnetic fluid for heat
transfer applications,” J. of Magnetism & Magnetic Mats., 252, pp. 360-362, 2002.
[138] Jones, P.D., “Radiation and convection heat transfer in particle-laden fluid flow,” Ph.D.
Thesis, Rice University, 1991.
[139] Weinstein, L.A., Loomis, J., Bhatia, B., Bierman, D.M., Wang, E.N. and Chen, G.,
“Concentrating solar power,” Chem. Rev., 115, pp. 12797-12838, 2015.
[140] Yong, S., Fu-Qiang, W., Xin-Lin, X., He-Ping, T., and Ying-Chun, L., “Radiative properties
of a solar cavity receiver/reactor with quartz window,” International journal of hydrogen
energy, 36, No. 19, pp. 12148-12158, 2011.

167
[141] Fleming, A., Ma, Z., Wendelin, T., Ban, H., and Folsom, C., “Thermal modeling of a multi-
cavity array receiver performance for concentrating solar power generation,” ASME 2015 9th
Int.l Conf. Energy Sus., ASME, 2015.
[142] Kim, H. N., Lee, H. J., Lee, S. N., Kim, J. K., Chai, K. K., Yoon, H. K., and Cho, H. S.,
“Experimental evaluation of the performance of solar receivers for compressed air,” J. of
Mech. Sci. & Tech., 28 No. 11, pp. 4789-4795, 2014.
[143] Ahn, Y., Lee, J., and Lee, J. I., “Preliminary studies on the heat exchanger option for S-CO
{sub 2} power conversion cycle coupled to water cooled SMR,” American Nuclear Society,
La Grange Park, IL, USA, 2012.
[144] Escandell, J., Wurm, D. J., Belleville, M. P., Sanchez, J., Harasek, M., and Paolucci-
Jeanjean, D., “Enzymatic synthesis of butyl acetate in a packed bed reactor under liquid and
supercritical conditions,” Catalysis Today, 255, pp. 3-9, 2015.
[145].Hesham G. Ibrahim, M.A.-M., Abubket A. Alshuiref, Abdulmajed R. Alagata and Mohamed
A. Adali, “Radial heat transfer in packed-beds-II: Mathematical modeling of heat transfer
through packed beds at elevated temperatures,” Int. J. of Eng. & App. Sci., 4, No. 2, pp. 24-
36, 2013.
[146]. Noorman, S., M. van Sint Annaland, and J.A.M. Kuipers, “Experimental validation of
packed bed chemical-looping combustion,” Chem. Eng. Sci., 65, No. 1, pp. 92-97, 2010.
[147]. Louge, M., J.M. Yusof, and J.T. Jenkins, “Heat transfer in the pneumatic transport of
massive particles,” Int. J. of Heat and Mass Transf., 1993. 36, No. 2, pp. 265-275, 1993.
[148]. A. Berlemont, O.S., M. Sommerfeld, “Validation of interparticle collision models based on
large Eddy simulation,” ASME/FED Symp. Fluids Eng. Div. Gas Particle Flows, 228, pp.
359-369, 1995.
[149]. Ahmadi, G. and D. Ma, “A thermodynamical formulation for dispersed multiphase turbulent
flows—1: Basic theory,” Int. J. of Multiphase Flow, 16, No. 2, pp. 323-340, 1990
[150]. Simonin, O., “Prediction of the dispersed phase turbulence in particle-laden jets,”
ASME/FED Symp. Fluids Eng. Div., Gas Solid Flows, 121, pp. 197, 1991.
[151]. J. Lavieville, E.D. and O. Simonin, “Large Eddy simulation of interactions between
colliding particles and a homogeneous isotropic turbulence field,” ASME/FED Symp. Fluids
Eng. Div., Gas Solid Flows, No. 228, pp. 347, 1995.
[152]. Cao, J. and G. Ahmadi, “Gas-particle two-phase turbulent flow in a vertical duct,” Int. J. of
Multiphase Flow,21, No.6, pp. 1203-1228, 1995.
[153]. Cao, J. and G. Ahmadi, “Gas–particle two-phase turbulent flow in horizontal and inclined
ducts,” Int. J. of Eng. Sci., 38, No.17, pp. 1961-1981, 2000.
[154]. Berlemont, A., P. Desjonqueres, and G. Gouesbet, “Particle lagrangian simulation in
turbulent flows,” Int. J. of Multiphase Flow, 16, No.1, pp. 19-34, 1990
[155]. T. Tanaka, Y.T., “Numerical simulation of gas-solid two-phase flow in a vertical pipe: on
the effect of inter-particle collision,” ASME/FED Symp. Fluids Eng. Div., Gas Solid Flows,
121, pp. 123, 1991.
[156]. Y. Yamamoto, M.P., T. Tanaka, T. Kajishima and Y. Tsuji, “Large-eddy simulation of
turbulent gas-particle flow in a vertical channel: effect of considering inter-particle
collisions,” J. of Fluid Mech., 442, pp. 303-334, 2001.
[157]. Zivkovic, M.S.a.G., “Recent advances in the numerical simulation of pneumatic conveying
through pipe systems,” Comp. Methods and App. Sci., Brussels, Belgium, pp. 201, 1992.
[158]. Oesterle, B. and A. Petitjean, “Simulation of particle-to-particle interactions in gas solid
flows,” Int. J. of Multiphase Flow, 19, No. 1, pp. 199-211, 1993.

168
[159]. Sommerfeld, M., “The importance of inter-particle collisions in horizontal gas-solid channel
flows,” ASME/FED Symp. Fluids Eng. Div., Gas Particle Flows, 228, pp. 335, 1995.
[160]. Lain, S., D. Bröder, and M. Sommerfeld, “Experimental and numerical studies of the
hydrodynamics in a bubble column,” Chem. Eng. Sci., 54, No. 21, pp. 4913-4920, 1999.
[161]. Manohar, A.A.A.a.K., “Theoretical and experimental investigation of heat transfer in
packed beds,” Research J. of App. Sci., 4, No. 5, pp. 166-177, 2009
[162]. Flamant, G., et al., “A New Heat Transfer Fluid for Concentrating Solar Systems: Particle
Flow in Tubes,” Energy Procedia, 49, No. 0, pp. 617-626, 2014.
[163] Mansoori, Z., Saffar-Avval, M., Tabrizi, H. B., Dabir, B., & Ahmadi, G. “Inter-particle heat
transfer in a riser of gas–solid turbulent flows,” Powder technology, 159, No. 1, pp. 35-45,
2005.
[164] Mansoori, Z., Saffar-Avval, M., Basirat-Tabrizi, H., Ahmadi, G., & Lain, S., “Thermo-
mechanical modeling of turbulent heat transfer in gas–solid flows including particle
collisions,” Int. J. of Heat & Fluid Flow, 23, No. 6, pp. 792-806, 2002.
[165] Mansoori, Z., Saffar-Avval, M., Tabrizi, H. B., and Ahmadi, G. “Experimental study of
turbulent gas–solid heat transfer at different particles temperature,” Exp. Therm. & Fluid Sci,
28, No.6, pp. 655-665, 2004.
[166] Awad, M. M., and Muzychka, Y. S. “Two-phase flow modeling in microchannels and
minichannels,” Heat Transfer Engineering, 31, No. 13, pp.1023-1033, 2010.
[167] Lim, K. S., Zhu, J. X., and Grace, J. R., “Hydrodynamics of gas-solid fluidization,”
International journal of multiphase flow, 21, pp.141-193, 1995.
[168] Rosenweig, R. E., “Fluidization: hydrodynamic stabilization with a magnetic field,” Science,
204, pp. 57-60, 1979.
[169] Kreith, F., Boehm, R.F., et al., “Heat and mass transfer,” Ch. 4, Mechanical Engineering
Handbook, Ed. Frank Kreith, Boca Raton, FL, USA, CRC Press LLC, 1999.
[170] Grace, J. R., “Contacting modes and behavior classification of gas-solid and other two-phase
suspensions,” Can. J. Chem. Eng. 64, pp. 353-363, 1986.
[171] Wen, C.Y. and Yu, Y.H., “A generalized method for predicting the minimum fluidization
velocity,” J. AIChE, 12, No. 2, pp. 610–612, 1966.
[172] Biyikli, K., Tuzla, K., and Chen, J.C., “Heat transfer around a horizontal tube in freeboard
region of fluidized beds,” J. AIChE, 29, No. 5, 712–716, 1983.
[173] Wen, C.Y. and MIller, E.N., “Heat transfer in solid-gas transport lines,” Ind. Eng. Chem.,
53, pp. 51–53, 1961.
[174] Lucas, A., Arnaldos, J., Casal, J. and Puigjaner, L., “Improved equation for the calculation
of minimum fluidization,” Ind. Eng. Chem. Proc. Des. Dev., 25, pp. 426-429, 1986.
[175] Adanez, J. & Abanades, J. C., “Minimum fluidization velocities of fluidized-bed
coalcombustion solids,” Powder Technol. 67, pp. 113-119, 1991.
[176] Chirone, R., Massimilla, L. and Russo, S., “Bubbling fluidization of a cohesive powder in
an acoustic field,” Fluidization VII, pp. 545-553. Engineering Foundation, New York, USA,
1992.
[177] Chirone, R., Massimilla, L. & Russo, S., “Bubble-free fluidization of a cohesive powder in
an acoustic field,” Chem. Eng. Sci. 48, pp. 41-52, 1993.
[178] Penchev, I. P. and Hristov, J. Y., “Behaviour of fluidized beds of ferromagnetic particles in
an axial magnetic field,” Powder Technol. 61, pp. 103-118, 1990.
[179] Bienert, W.B., Rind, H., Wolf, A.A., “Conceptual design of an open cycle air Brayton Solar
Receiver: Phase 1 final report,” DTM-79, contract no. 955135 for California Institute of

169
 Technology Jet Propulsion Laboratory, Dynatherm Corporation, Cockeysville, MD, USA,
1979.
[180] Heller, P., “Solar-hybrid power and cogeneration plants” SOLHYCO, Project No. 019830,
2011.
[181] Heller, P., Pfander, M., Denk, T., Tellez, F., Valverde, A., Fernandez J, et al., “Test and
evaluation of a solar powered gas turbine system,” Solar Energy, 80, No. 10, pp.1225–1230,
2006.
[182] Amsbeck, L., Buck, R., Heller, P., Jedamski J, and Uhlig R., “Development of a tube receiver
for a solar-hybrid microturbine system,” SolarPACES 2008, Las Vegas, NV, USA, 2008
[183] Amsbeck, L., Denk, T., Ebert, M., Gertig C, Heller P, Herrmann P, et al., “Test of a solar-
hybrid microturbine system and evaluation of storage deployment,” SolarPACES 2010,
Perpignan, France, 2010.
[184] Uhlig, R., “Transient stresses at metallic solar tube receivers,” SolarPACES 2011, Granada,
Spain, 2011
[185] Kolb, G.J., “An evaluation of possible next-generation high-temperature molten-salt power
towers,” Sandia National Laboratories, SAND2011-9320, Albuquerque, NM, USA 2011.
[186] Becker, M., and Boehmer, M., "GAST: The Gas Cooled Solar Tower Technology Program,"
Springer, Berlin, Germany, 1989.
[187] Romero-Alvarez, M., & Zarza, E., “Concentrating solar thermal power,” Handbook of
energy efficiency and renewable energy, Ch. 21, 2007.
[188] Hennecke K, Schwarzbozi P, Alexopoulos S, Gottsche J, Hoffschmidt B, Beuter M, et al.,
“Solar power tower Julich: the first test and demonstration plant for open volumetric receiver
technology in Germany,” SolarPACES 2008, Las Vegas, NV, 2008.
[189] Avila-Marin, A.L., “Volumetric receivers in solar thermal power plants with central receiver
system technology: a review,” Solar Energy, 85, No. 5, pp. 891-910, 2011.
[190] Zunft S., Hanel M., Kruger M, Dreissigacker V., Gohring F., and Wahl E. Julich, “Solar
power tower-experimental evaluation of the storage subsystem and performance calculation,”
J. of Sol. Energy Eng.-Trans. ASME, 133, No. 3, 2011.
[191] Menigault T., Flamant G., and Rivoire B., “Advanced high-temperature 2-slab selective
volumetric receiver,” Solar Energy Materials, 24, pp.192–203, 1991.
[192] Pitzpaal R., Morhenne J. and Fiebig M.A., “New concept of a selective solar receiver for
high-temperature applications,” Solar Energy Materials, 24, pp. 293–306, 1991.
[193] Kribus, A., Ries, H. and Spirkl, W., “Inherent limitations of volumetric solar receivers,” J.
of Sol. Energy Eng.-Trans. ASME, 118 No. 3, 1996.
[194] Karni, J., Kribus, A., Rubin, R. and Doron, P., “The porcupine: a novel high-flux absorber
for volumetric solar receivers,” J. of Solar Energy Eng.-Trans. ASME, 120, No. 2, 1998.
[195] Modi, A., Knudsen, T., Haglind, F., and Clausen, L. R., “Feasibility of using ammonia-water
mixture in high temperature concentrated solar power plants with direct vapour generation,”
Energy Procedia, 57, pp. 391-400, 2014.
[196] Becker, M., “Comparison of heat transfer fluids for use in solar thermal power stations,”
Electric Power Systems Research, 3, pp. 139-150, 1980.
[197] Heller, L., “Literature review on heat transfer fluids and thermal energy storage systems in
CSP plants,” STERG Report, 2013.
[198] Invernizzi, C.M., “The thermodynamic properties of working fluids,” Ch. 2, Lecture Notes
in Energy 11, Springer-Verlag, London Ltd., 2013.

170
[199] Pitzer, K.S., Lippmann, D.Z., Curl, R.F., Huggins, C.M., and Petersen, D.E., “The
volumetric and thermodynamic properties of fluids. II. Compressibility factor, vapour
pressure and entropy of vaporization,” J. Am. Chem. Soc., 77, No. 13, pp. 3433–3440, 1955.
[200] NREL, “Concentrating solar power projects,”, 2015, http://www:nrel :gov/ csp/ solarpaces/
by technology:cfm.
[201] Dow Chemical Company, “`Synthetic Organic Heat Transfer Fluid Liquid and Vapor Phase
Data,” 2015, http://www:dow:com/ webapps/ lit/ litorder:asp? _lepath=heattrans/ pdfs/noreg/
176-01463 :pdf&pdf=true.
[202] Pacheco, J.E., Moursund, C., Rogers, D., Wasyluk, D., “Conceptual design of a 100MWe
modular molten salt power tower plant,” SolarPACES2011, Granada, Spain, 2011.
[203] Dunham, M. T., and Iverson, B. D., “High-efficiency thermodynamic power cycles for
concentrated solar power systems,” Renewable and Sustainable Energy Reviews, 30, pp. 758-
770, 2014.
[204] Pacheco, J. E., Reilly, H. E. and Gregory, J., “Summary of the solar two test and evaluation
program, Technical report, Sandia National Laboratories, SAND2000-0372C, 2000.
[205] Yang, Z. and Garimella, S. V., “Molten-salt thermal energy storage in thermoclines under
different environmental boundary conditions,” Applied Energy, 87, No. 11, pp. 3322-3329,
2010.
[206] Garbrecht, O., Al-sibai, F., Kneer, R. and Wieghardt, K., “Numerical Investigation of a new
molten salt central receiver design,” SolarPACES 2012, Marrakech, Morocco, 2012.
[207] Pacheco, J. E. and Dunkin, S. R., “Assessment of Molten-Salt Solar Central- Receiver
Freeze-Up and Recovery Events,” Technical report SAND96-0331C, Sandia National
Laboratories, Albuquerque, USA, 1996.
[208] Kelly, B., “Advanced Thermal Storage for Central Receivers with Supercritical Coolants,”
Technical report DE-FG36-08GO18149, Abengoa Solar, Lakewood, Colorado, 2010.
[209] Raade, J. W. and Padowitz, D., “Development of Molten Salt Heat Transfer Fluid with Low
Melting Point and High Thermal Stability,” J. of Solar Energy Eng. 133, No. 3, pp. 31013-
31016, 2011.
[210] Das, A. K., Clark, M. M., Teigen, B. C., Fiveland, W. A., and Anderson, M. H., “Heat
transfer behavior of molten nitrate salt,” SOLARPACES 2015, 1734, No. 1, 2016.
[211] J.W. Cooke and B. Cox, “Forced-Convection Heat-Transfer Measurements with a Molten
Fluoride Salt Mixture Flowing in a Smooth Tube,” Oak Ridge National Lab, 1973.
[212] Silverman, M.D., Huntley, W.R. and Robertson, H.E., “Heat transfer measurements in
forced convection loop with two molten-fluoride salts: LiF-BeF2-ThF4-UF4 and eutectic
NaBF4-NaF,” Oak Ridge National Lab, 1976.
[213] Hoffman, H.W., “Turbulent forced convection heat transfer in circular tubes containing
molten sodium hydroxide,” Oak Ridge National Lab, 1952.
[214] Grele, M.D., Gedeon, L., “Forced-convection heat-transfer characteristics of molten sodium
hydroxide,” NACA, 1953.
[215] Hoffman, H.W., Cohen, S.I., “Fused Salt Heat Transfer - Part III: Forced-Convection Heat
Transfer in Circular Tubes Containing the Salt Mixture NaNO2-NaNO3-KNO3,” Oak Ridge
National Lab, 1960.
[216] W. Yu-ting, L. Bin, M. Chong-fang, G. Hang, “Convective heat transfer in the laminar-
turbulent transition region with molten salt in a circular tube,” Exp. Therm. Fluid Sci., 33, pp.
1128–1132, 2009.

171
[217] Gnielinski, V., “New equations for heat and mass-transfer in turbulent pipe and channel
flow,” Int. chem. Eng., 16, No. 2, pp. 359-368, 1976.
[218] Andraka, C.E., Rawlinson, K.S., Moss, T.A., Adkins, D.R., Moreno, J.B. and Gallup, D.R.,
“Solar Heat Pipe Testing of the Stirling Thermal Motors 4-120 Stirling Engine,” IECEC 96,
1292, pp. 1295-1300, 1996.
[219] Andraka, C.E., Rawlingson, K.S. and Siegel, N.P., “Technical feasibility of storage on large
dish Stirling systems,” Sandia Report SAND2012-8352, Sandia National Laboratories,
Albuquerque, NM, USA, 2012.
[220] Angelino, G., and Invernizzi, C., “Binary conversion cycles for concentrating solar power
technology,” Solar Energy, 82, No. 7, pp.637-647, 2008.
[221] Goswami, D.Y., Kreith, F., Kreider, J.F., “Principles of solar engineering,” 2nd Ed., Taylor
& Francis, Philadelphia, PA, USA, 1999.
[222] Boerema, N., Morrison, G., Taylor, R. and Rosengarten, G.,“Liquid sodium versus Hitec as
a heat transfer fluid in solar thermal central receiver systems,” Solar Energy, 86, No. 9, pp.
2293-2305, 2012.
[223] Lata, J. M., Rodriguez, M. and de Lara, M. A., “High Flux Central Receivers of Molten Salts
for the New Generation of Commercial Stand-Alone Solar Power Plants,” J. of Solar Energy
Eng. 130, No. 2, 2008.
[224] Coventry, J., Pye, J., Andraka, C.E., Blanco, M. and Fisher, J., “Sodium receivers for solar
power towers: a review,” Energy Procedia, 00, 2015.
[225] J.P. Kotze, T.W. von Backstrom and P.J. Erens, “NaK as a Primary Heat Transfer Fluid in
Thermal Solar Power Installation,” SASEC 2012, South Africa, 2012.
[226] Armijo, K. M., and Andraka, C. E. “Phenomenological studies on sodium for CSP
applications: A safety review,” SOLARPACES 2015, 1734, No. 1, 2016.
[227] Morozov, Yu.D., Prisniakov, V.F., Privalov, A.N., Trofimenko, A.V., Gabrinets, V.A.,
Gontarev, Yu.K., “Solar dynamic Rankine space power plant using vapour–liquid metallic
techniques,” Space Technology, 16, No. 4, pp. 169–176, 1996.
[228] Fraas, A.P., “A potassium-steam binary vapor cycle for better fuel economy and reduced
thermal pollution,” Trans. ASME J. of Energy & Power, pp. 53–63, 1973.
[229] Foust, O. J., “Sodium-NaK Engineering Handbook,” Gordon and Breach, Science
Publishers Inc., New York, USA, 1972.
[230] Diver, R. B., Adkins, D. R. and Moss, T. A., “Trends in Dish-Stirling Solar Receiver
Designs,” Energy Conv. Eng. Conf., pp. 303-310, 1990.
[231] Chopra, O.K., Smith, D.L., Tortorelli, P.F., DeVan, J.H., and Kai Sze, D., “Liquid–metal
corrosion,” Fusion Technology, 8, pp. 1956–1969, 1985.
[232] Song, J.S., Louis, J.F. and Wendte, J.C., “Liquid lithium thermal energy storage for solar
dynamic power systems,” Proc. of 23rd Intersociety Energy Conv. Eng. Conf., Denver, CO,
USA, 2, pp. 187–193, 1988.
[233] Davis, J.P. and Kikin, G.M., “Lithium-boiling potassium Rankine cycle test operating
experience,” Adv. in Energy Conv. Eng., Intersoc. Energy Conv. Eng. Conf., Miami Beach,
Florida, USA, pp. 87–95, 1967.
[234] International Energy Agency, “Central Receiver System (CRS),” Small Solar Power
Systems Project (SSPS), 1, Springer, USA, 1986.
[235] International Energy Agency, “Book of Summaries,” 4, Springer, USA 1986.

172
[236] D.R. Adkins, C.E. Andraka, J.B Moreno, K.S. Rawlinson, S.K. Showalter and T.A Moss,
“Heat pipe solar receiver development activities at Sandia National Laboratories,” Renewable
and Advanced Energy Conference for the 21st Century, Maui, Hi, 1999.
[237] Olivier, T.J., Radel, R.F., Nowlen, S.P., Blanchat, T.K. and Hewson, J.D., “Metal fire
implications for advanced reactors, Part 1: Literature review,” Sandia Report, SAND2007-
6332, Sandia National Laboratories, Albuquerque, NM, USA, 2007.
[238] Griffin, C.W. and Piccot, A.R., “Sodium-NaK Engineering handbook – Sodium Flow, Heat
Transfer, Intermediate Heat Exchangers, and Steam Generators,” Ch. 2, 3, Gordon and Breach
Inc., New York, 1976.
[239] Wegrzyn, J. and Jhaw, D.T., “Experimental investigation of aerosol behavior in Post
LMFBR accident reactor containment atmosphere,” NUREG/CR-0799.R7-38.45, U.S.
Nuclear Regulatory Commission, 1979.
[240] Baqué, F. and Beauchamp, F., “Sodium – Water reaction in steam generators,” ESFR SP5,
2009.
[241] Schulenberg, T., Cheng, X., and Stieglitz, R., “Thermal-hydraulics of Lead-Bismuth for
Accelerator Driven Systems,” NURETH11, Avignon, France, 2005.
[242] Fazio, C., Sobolev, V. P., Aerts, A., Gavrilov, S., Lambrinou, K., Schuurmans, P. and Gosse,
S. “Handbook on Lead-bismuth Eutectic Alloy and Lead Properties, Materials Compatibility,
Thermal-hydraulics and Technologies-2015 Edition” No. NEA--7268, 2015.
[243] Kim, J. S., Dawson, A., Wilson, R., Venkatesan, K., and Stein, W., “High-Temperature Heat
Transport and Storage Using LBE Alloy for Concentrated Solar Power System,” ASME 2015
9th Int. Conf. on Energy Sus., 2015.
[244] Pacio, J. and Wetzel, T., “Assessment of liquid metal technology status and research paths
for their use as efficient heat transfer fluids in solar central receiver systems,” Solar Energy,
93, pp. 11-22, 2013.
[245] Hennecke, K., Schwarzbözl, P., Hoffschmidt, B., Göttsche, J., Koll, G. and Beuter, M., “The
Solar Power Tower Jülich: A solar thermal power plant for test and demonstration of air
receiver technology,” Proc. of ISES World Congress, 2009.
[246] Koll, G., Sahraoui, T., Höschmidt, B., Khedim, A., Pomp, S., Schwarzbözl, P. and Dillig,
M., “ALSOL Solar Thermal Tower Power Plant Algeria,” SolarPACES 2011, Ganada, Spain
2011.
[247] Wilson Solarpower, “Wilson’s concentrated solar power (CSP) system targets 5-6
cents/kilowatt-hour,” 2015, http://wilsonsolarpower.com/
[248] Frutschi, H.U., “Closed-cycle Gas Turbines,” Operating Experience and Future Potential,”
ASME Press, New York, USA, pp. 200–228, 2005.
[249] Tsiklauri, G.,Talbert, R., Schmitt, S., Filippov, G., Bogoyavlensky, R., Grishanin, E, et al.,
“Supercritical steam cycle for nuclear power plant,” Nucl. Eng. Des., 235, 2005.
[250] Forsberg, C.W., Peterson P.F. and Zhao, H. “High-temperature liquid-fluoride-salt closed-
Brayton-cycle solar power towers,” J. Sol. Energy Eng., 129, 2007.
[251] Fleming, D., Kruizenga, A., Pasch, J., Conboy, T., and Carlson, M., “Corrosion and Erosion
Behavior in Supercritical CO2 Power Cycles,” ASME Turbo Expo 2014: Turbine Technical
Conference and Exposition, 2014.
[252] Boss, M., Gadoury, T., Feeny, S. and Montgomery, M., “Recent Advances in Ultra Super
Critical Steam Turbine Technology,” Technical report, 2007.
[253] Weitzel, P. S., “Steam Generator for Advanced Ultra-Supercritical Power Plants 700 to
760C,” ASME 2011 Power Conference, Denver, CO, USA, 2011.

173
[254] Coventry, J. and Pye, J., “Coupling Supercritical and Superheated Direct Steam Generation
with Thermal Energy Storage,” Technical report, Wizard Power, 2010.
[255] Ahn, Y., Bae, S. J., Kim, M., Cho, S. K., Baik, S., Lee, J. I., and Cha, J. E., “Review of
supercritical CO2 power cycle technology and current status of research and development,”
Nuclear Engineering and Technology, 47, No. 6, pp. 647-661, 2015.
[256] Persichilli, M., Kacludis, A., Zdankiewicz, E., and Held, T., “Supercritical CO2 Power Cycle
Developments and Commercialization: Why sCO2 can Displace Steam Ste,” Power-Gen
India & Central Asia, 2012.
[257] Wright, S. A., Conboy, T. M., Parma, E. J., Tom G. Lewis, Gary A. Rochau and Ahti J. Suo-
Anttila, “Summary of the Sandia Supercritical CO2 Development Program,” SCO2 Power
Cycle Symposium, Boulder, CO, 2011.
[258] Turchi, C. S., “Supercritical CO2 for Application in Concentrating Solar Power Systems,”
Proc. of SCO2 Power Cycle Symp., Troy, NY, USA, 2009.
[259] Gary, J., Turchi, C. and Siegel, N., “CSP and the DOE SunShot Initiative,” SolarPACES
2011, 1, Granada, Spain, 2011.
[260] Chapman, D. J. and Arias, D. A., “An Assessment of the Supercritical Carbon Dioxide Cycle
for Use in a Solar Parabolic Trough Power Plant,” Proc. of SCO2 Power Cycle Symp., Troy,
NY, USA, 2009.
[261] Ma, Z. and Turchi, C. S., “Advanced Supercritical Carbon Dioxide Power Cycle
Configurations for Use in Concentrating Solar Power Systems,” Supercritical CO2 Power
Cycle Symp., Boulder, CO, 2011.
[262] Cox, T. L., “Preliminary Results of Heat Transfer Experiments with Supercritical Carbon
Dioxide in the Small System Test Loop,” Proc. of SCO2 Power Cycle Symp., Troy, NY, 2009.
[263] Rouillard, F., Charton, F. and Moine, G., “Corrosion behaviour of different metallic
materials in supercritical CO2 at 550 C and 250 bars,” Proc. of SCO2 Power Cycle Symp.,
Troy, NY, 2009.
[264] Dostal, V., Driscoll, M. J. and Hejzlar, P., “A Supercritical Carbon Dioxide Cycle for Next
Generation Nuclear Reactors,” Technical report, Massachusetts Institute of Technology,
Cambridge, MA., 2004.
[265] Spiers, D.J., and Rasinkoski, A.D., “Predicting the service lifetime of lead/acid batteries in
photovoltaic systems,” J. of Power Sources, 53, No. 2, pp. 245–253, 1995.
[266] Li, P. W., Chan, C. L., Stephens, J., James. E. O'Brien, Van Lew, J., and Karaki, W.,
“Transient heat transfer and energy transport in packed bed thermal storage systems,”
INTECH Open Access Publisher, 2011.
[267] Herrmann, U. and Kearney D.W., “Survey of thermal energy storage for parabolic trough
power plants,” J. of Sol. Energy Eng., 124, No. 2, pp. 145–152, 2002.
[268] Kuravi, S., Trahan, J., Goswami, D. Y., Rahman, M. M., & Stefanakos, E. K., “Thermal
energy storage technologies and systems for concentrating solar power plants,” Prog. in
Energy & Comb. Sci., 39, No. 4, pp. 285-319, 2013.
[269] Sioshansi, R. and Denholm, P., “The value of concentrating solar power and thermal energy
storage,” NREL-TP-6A2-45833, NREL, Golden, CO, USA, 2010.
[270] Reilly, H.E. and Kolb, W.J., “Evaluation of molten salt power tower technology based on
the experience of solar two,” SAND2001-3674, Sandia National Laboratories, Albuquerque,
NM, USA, 2001.
[272] Jorgenson, J., Denholm, P., and Mehos, M., “Quantifying the Value of Concentrating Solar
Power in a Production Cost Model,” ASME 2014 8th Int. Conf. on Energy Sus., 2014.

174
[272] Sioshansi, R., and Denholm, P., “The value of concentrating solar power and thermal energy
storage,” IEEE Transactions on Sustainable Energy, 1, No. 3, pp. 173-183, 2010.
[273] CPUC, “2015 Filing Guide for System, Local and Flexible Resource Adequacy (RA)
Compliance Filings,” Issued: September 9, 2014, 2014.
[274] Mehos, M., Turchi, C., Jorgenson, J., Denholm, P., Ho, C., and Armijo, K., “On the Path to
SunShot: Advancing Concentrating Solar Power Technology, Performance, and
Dispatchability,” No. NREL/TP-5500-65688; SAND2016-2237, NREL, Golden, CO, USA,
2016.
[275] Miller, N.W., M. Shao, S., Pajic, and R. D'Aquila, “Western Wind and Solar Integration
study phase 3-Frequency Response and Transient Stability,” Technical Report NREL/SR-
5D00-62906, 2014.
[276] Madaeni, S. H., Sioshansi, R., and Denholm, P., “Capacity value of concentrating solar
power plants,” Contract, 303, pp. 275-3000, 2011.
[277] Mills, A., and R. Wiser, “An Evaluation of Solar Valuation Methods Used in Utility Planning
and Procurement Processes,” Environmental Energy Technologies Division, LBNL, LBNL‐
5933E, 2012.
[278] Michels, H., and Pitz-Paal, R., “Cascaded latent heat storage for parabolic trough solar power
plants,” Solar Energy, 81, No. 6, pp. 829-837, 2007.
[279] Ma, Z., Glatzmaier, G. C., Wagner, M., and Neises, T., “General Performance Metrics and
Applications to Evaluate Various Thermal Energy Storage Technologies,” ASME 2012 6th
Int. Conf. on Energy Sus., pp. 345-351, 2012.
[280] Zhao, C.Y., Wu, C.G., “Thermal property characterization of a low melting-temperature
ternary nitrate salt mixture for thermal energy storage systems,” Sol. Energy Mater. Sol. Cells,
95, pp. 3341–3346, 2011.
[281] Herrmann U. and Kearney D.W., “Survey of thermal energy storage for parabolic trough
power plants,” Solar Energy Engineering, 124, pp. 145-152, 2002.
[282] Baumann, T., and Zunft, S., “Properties of granular materials as heat transfer and storage
medium in CSP application,” Solar Energy Materials and Solar Cells, 143, 38-47, 2015.
[283] Seyf, H. R., and Henry, A., “Thermophotovoltaics: a potential pathway to high efficiency
concentrated solar power,” Energy & Env. Sci., 9, No. 8, pp. 2654-2665, 2016.
[284] H Abedin, A., and A Rosen, M., “A critical review of thermochemical energy storage
systems,” The Open Renewable Energy Journal, 4, No. 1, 2011.
[285] Lacy, D.E., Coles-Hamilton, C. and Juhasz, A., “Selection of high temperature thermal
energy storage materials for advanced solar dynamic space power systems,” NASA Technical
Memorandum 89886, 22nd Int. Energy Conv. Eng. Conf., Philadelphia, PA, USA, 1987.
[286] Wallin, W.E. and Dustin, M.O., “Advanced space solar dynamic power systems beyond IOC
space station,” Energy New Frontiers, 22nd Intersoc. Energy Conv. Eng. Conf., 1,
Philadelphia, PA, USA, pp. 145–155, 1987.
[287] Kesseli, J.B. and Lacy, D.E., “Advanced solar receiver conceptual design study,” Energy
New Frontiers, 22nd Intersoc. Energy Conv. Eng. Conf., Philadelphia, PA, USA, 1, pp. 162–
168, 1987.
[288] Misra, A.K. and Whittenberg, J.D., “Fluoride salts and container materials for thermal
energy storage applications in the temperature range 973–1400 K,” Energy New Frontiers,
22nd Intersoc. Energy Conv. Eng. Conf., Philadelphia, PA, USA, 1, pp. 188–201, 1987.

175
[289] Shabgard, H., Bergman, T. L., and Faghri, A., “Exergy analysis of latent heat thermal energy
storage for solar power generation accounting for constraints imposed by long-term operation
and the solar day,” Energy, 60, pp. 474-484, 2013.
[290] Regin, A. F., Solanki, S. C., and Saini, J. S., “Heat transfer characteristics of thermal energy
storage system using PCM capsules: a review,” Renewable and Sustainable Energy Reviews,
12, No. 9, pp. 2438-2458, 2008.
[291] Nithyanandam, K., and Pitchumani, R., “Cost and performance analysis of concentrating
solar power systems with integrated latent thermal energy storage. Energy,” 64, pp. 793-810,
2014.
[292] Andraka, C. E., Kruizenga, A. M., Hernandez-Sanchez, B. A., and Coker, E. N., “Metallic
phase change material thermal storage for dish Stirling,” Energy Procedia, 69, pp. 726-736,
2015.
[293] Petrone, G., Spagnuolo, G., Teodorescu, R., Veerachary, M., and Vitelli, M., “Reliability
issues in photovoltaic power processing systems,” IEEE Transactions on Industrial
Electronics, 55, No. 7, pp. 2569-2580, 2008.
[294] Weber, R. and Dorer, V. “Long-term heat storage with NaOH,” Vacuum, 82, No. 7, pp. 708-
716, 2008.
[295] Mouromtseff, I.E., “Water and forced-air cooling of vacuum tubes nonelectronic problems
in electronic tubes,” Proc. IRE, 40, No. 4 pp. 190-205, 1942.
[296] Bonilla, C., Baumeister, T. and Dunning, J.R., “Nuclear Engineering,” McGraw-Hill, New
York, USA, 1975.
[297] Lenert, A., Nam, Y. and Wang, E.N., “Heat transfer fluids,” Annu. Rev. Heat Transfer, 15,
pp. 93-129, 2012.
[298] Bales, C., Setterwall, F. and Bolin G., “Development of the thermo-chemical accumulator
(TCA),” Proc. of Eurosun, Freiburg, Germany, 2004.
[299] Mauran, S., Lahmidi, H. and Goetz, V. “Solar heating and cooling by a thermochemical
process: first experiments of a prototype storing 60 kWh by a solid/gas reaction,” Solar
Energy, 82, No. 7, pp. 623-636, 2008.
[300] Zondag, H., van Essen, M., He, Z., Schuitema, R. and van Heldan, W., “Characterization of
MgSO4for Thermochemical Storage,” 2nd Int. Ren. Energy Storage Conf. (IRES II), Bonn,
Germany, 2007.
[301] Azpiazu, M.N., Morquillas, J.M. and Vazquez, A. “Heat recovery from a thermal energy
storage based on the Ca(OH)2/CaO cycle,” Appl. Therm. Eng., 23, No. 6, pp. 733-741, 2003.
[302] Lovergrove, K., Luzzi, A. and Kreetz, H. “A solar-driven ammonia-based thermochemical
energy storage system,” Solar Energy, 67, No. 4, pp. 309-316, 1999.
[303] Kreetz, H. and Lovergrove, K. “Theoretical analysis and experimental results of a 1 kWchem
ammonia synthesis reactor for a solar thermo-chemical energy storage system,” Solar Energy,
67, No. 4, pp. 287-296, 1999.
[304] Winter, C. J., Sizmann, R. L., and Vant-Hull, L. L., “Solar Power Plants”, Springer, Berlin,
Germany, pp. 21–27. 1991.
[305] Kodama, T. “High-temperature solar chemistry for converting solar heat to chemical fuels,”
Prog. in Energy and Comb. Sci., 29, pp. 567–597, 2003.
[306] Seinfeld, A., “Solar thermochemical production of hydrogen-a review,” Solar Energy, 78,
pp. 603–615, 2005.
[307] Finch, N.S. and Ho, C.K., “Stochastic modeling of power tower sand evaluation of technical
improvement opportunities,” SolarPACES 2011. Granada, Spain, 2011.

176
[308] Bolland O., “A comparative evaluation of advanced combined cycle alternatives,” Proc. of
Int. Gas Turbine & Aeroengine Congress & Exhibition, Brussels, Belgium, 1990.
[309] Mittelman, G., Epstein, M., “A novel power block for CSP systems,” SolEnergy, 84, pp.
1761–71, 2010.
[310] Spelling, J., Favrat, D., Martin, A., Augsburger, G., “Thermoeconomic optimization of a
combined-cycle solar tower power plant,” Energy, 41 pp.113–120, 2012.
[311] Glatzmaier, G., “Summary Report for Concentrating Solar Power Thermal Storage
Workshop: New Concepts and Materials for Thermal Energy Storage and Heat-Transfer
Fluids,” NREL/TP-5500-52134, NREL, Golden, CO., 2011.
[312] Haywood, R.W., “Analysis of engineering cycles,” 3rd Ed., New York, USA, Pergamon
Press, 1991.
[313] Kehlhofer, R., Bachmann, R., Nielson, H., Warner, J., “Combined-cycle gas and steam
turbine power plants,” 2nd Ed., Tulsa, OK, USA, Pennwell Publishing; 1999.
[314] McGovern, R.K., Smith, W.J., “Optimal concentration and temperatures of solar thermal
powerplants,” Energy Conver. Manag., 60 pp. 226–232, 2012.
[315] Omer, L. M., and Mustafa, N. M., “Performance Of Laboratory Small Scale Power Plant
Based On Rankine Cycle For Rural Area Power Generation,” Ph.D. Dissertation, University
of Khartoum, 2015.
[316] Kribus, A., Zaibel, R., Carey, D., Segal, A., and Karni, J., “A solar-driven combined cycle
power plant,” Solar energy, 62, No. 2, pp. 121-129, 1998.
[317] Barigozzi, G., Bonetti, G., Franchini, G., Perdichizzi, A., Ravelli, S., “Thermal performance
prediction of a solar hybrid gas turbine,” SolEnergy, 86, pp. 2116–2127, 2012.
[318] Price, H., Lupfert, E., Kearney, D., Zarza, E., Cohen, G., Gee, R., et al., “Advances in
parabolic trough solar power technology,” SolEnergy Eng., 124, pp. 109–125, 2002.
[319] Zarza, E., Rojas, M.E., Gonzalez, L., Caballero, J.M., Rueda, F., “INDITEP: The first pre-
commercial DSG solar power plant,” SolEnergy, 80, pp. 1270–1276, 2006.
[320] Wheeldon, J.M., “Engineering and economic evaluation of 1300ᵒF series ultra- supercritical
pulverized coal power plants: phase 1,” No. 1015699. PaloAlto, CA, USA, Electric Power
Research Institute, 2008.
[321] Chen, H., Goswami, D.Y., Stefanakos, E.K., “A review of thermodynamic cycles and
working fluids for the conversion of low-grade heat,” Renew. Sustain. Energy, 14 pp. 3059–
3067, 2010.
[322] Fernández, F.J., Prieto, M.M., Suárez, I., “Thermodynamic analysis of high-temperature
generative organic Rankine cycles using siloxanes as working fluids,” Energy, 36, pp. 5239-
5249, 2011.
[323] Yeo, W.H., Fry, A.T., Purbolaksono, J., Ramesh, S., Inayat-Hussain, J.I., Liew, H.L., and
Hamdi, M., “Oxide scale growth and presumed exfoliation in a 700° C or higher steam
condition: A simulation study for future operations of ultra-supercritical power plants,”The J.
of Supercritical Fluids, 92, pp. 215-222, 2014.
[324] Ries, H., Kribus, A., Karni, J., “Non-isothermal receivers,” J. SolEnergy Eng., 117, pp. pp.
259–261, 1995.
[325]Thermonet, “Brayton Cycle Irreversibilities,” Wiley, 2001,
http://www.wiley.com/college/moran/CL_0471465704_S/user/tutorials/tutorial9/tut9m_cont
ent.html

177
[326] Greitzer, E.M., Spakovszky, Z.S. and Waitz, I.A., “Unified: Thermodynamics and
propulsion,” Ch. 16 Lecture Notes, MIT, Accessed: Aug. 2016:
http://web.mit.edu/16.unified/www/FALL/thermodynamics/notes/notes.html
[327] Paul, S., Shepherd, R., & Woollin, P. (2012). Selection of materials for high pressure CO2
transport. TWI. 2012, http://www.twi.com.uk.
[328] Ahn, Y., Bae, S. J., Kim, M., Cho, S. K., Baik, S., Lee, J. I., and Cha, J. E., “Review of
supercritical CO2 power cycle technology and current status of research and
development,” Nuclear Engineering and Technology, 47, No. 6, 647-661, 2015.
[329] Hill, P. G., and Peterson, C. R., “Mechanics and thermodynamics of propulsion,” Addison-
Wesley Publishing Co., 764, 1992.
[330] J. C. Bryant, H. Saari, and K. Zanganeh, “An Analysis and Comparison of the Simple and
Recompression Supercritical CO2 Cycles,” Supercritical CO2 Power Cycle Symposium,
2011.
[331] Chen, Y., “Novel cycles using carbon dioxide as working fluid,” Licentiate Thesis,
Universitetsservice US AB, Stockholm, Sweden, 2006
[332] Goswami, D.Y., “Thermodynamic cycles for CSP,” Clean Energy Center Presentation,
Univ. of South Florida, 2015.
[333] Dostal, V., Driscoll, M.J. and Hejzlar, P., “A Supercritical Carbon Dioxide Cycle for Next
Generation Nuclear Reactors,” MITANP-TR-100, MIT, Cambridge (MA), 2004. Available
from: http://hdl.handle.net/1721.1/17746.
[334] Moisseytsev, A. and Sienicki, J.J., “Investigation of alternative layouts for the supercritical
carbon dioxide Brayton cycle for a sodium-cooled fast reactor,” Nucl. Eng. Des., 239, pp.
1362–1371, 2009.
[335] Iverson, B.D., Conboy, T.M., Pasch, J.J. and Kruizenga, A.M., “Supercritical CO2 Brayton
cycles for solar-thermal energy,” Appl. Energy, 111, pp. 957–970, 2013.
[336] Garg, P., Kumar, P. and Srinivasan, K., “Supercritical carbon dioxide Brayton cycle for
concentrated solar power,” J. of Supercrit. Fluids, 76, pp. 54–60, 2013.
[337] Singh, R., Rowlands, A.S., Miller, S.a., “Effects of relative volume-ratios on dynamic
performance of a direct-heated supercritical carbon-dioxide closed Brayton cycle in a solar-
thermal power plant,” Energy, 55, pp. 1025–1032, 2013.
[338] Persichilli, M., Kacludis, A., Zdankiewicz, E. and Held, T. “Supercritical CO2 Power Cycle
Developments and Commercialization: Why sCO2 can Displace Steam,” Ste. Power-Gen
India & Central Asia, 2012.
[339] Ames, R., Driscoll, D. and Narayanan, V., “Design, Fabrication and Characterization of
Microchannel Heat Exchangers for Fossil-Fired Supercritical CO2 Cycles,” DOE Project
FE0024064, NETL, 2016, https://www.netl.doe.gov/research/coal/project-
information/proj?k=FE0024064
[340] Neises, T. and Turchi, C., “A comparison of supercritical carbon dioxide power cycle
configurations with an emphasis on CSP applications,” Energy Procedia, 49, pp. 1187-1196,
2014.
[341] Forsberg, C. W., Peterson, P. F. and Zhao, H., “High-Temperature LiquidFluoride-Salt
Closed-Brayton-Cycle Solar Power Towers,” “J. of Solar Energy Eng. 129, No. 2, 2007.
[342] Dunham, M.T. and Lipinski, W., “Thermodynamic analyses of single Brayton and combined
Brayton-Rankinecycles for distributed solar thermal power generation,” J. of Sol. Energy
Eng., 135, 2013.

178
[343] Chacartegui, R., Munoz de Escalona, J.M., Sanchez, D., Monje, B. and Sanchez, T.,
“Alternative cycles based on carbon dioxide for central receiver solar power plants,” Appl.
Therm. Eng., 31, pp. 872-879, 2011.
[344] Johnson, G.A., “Power conversion system evaluation for next generation nuclear plant,”
Nuclear, Tech., 175, pp. 371-387, 2011.
[345] Sabharwall, P., Kim, E. S., Mckellar, M. and Patterson, M., “Small Modular Molten Salt
Reactor (SM-MSR),” ASME 2011 Small Modular Reactors Symp., pp. 31-39, 2011
[346] B-Cubed, “The Brayton Cycle,” Accessed Aug. 2016, http://www.learnthermo.com/
[347] Bryant, J. C., Saari, H., and Zanganeh, K., “An analysis and comparison of the simple and
recompression supercritical CO2 cycles,” Supercritical CO2 Power Cycle Symposium,
Boulder, CO, 2011.
[348] Kulhanek, M., & Dostal, V., “Supercritical carbon dioxide cycles thermodynamic analysis
and comparison,” Supercritical CO2 Power Cycle Symposium, Boulder, CO, 2011.
[349] Sugarmen, C., Tellez, F., Buck, R., Medina, J. E., Altwegg, P., Ring, A., and Schwarzbözl,
P., “SOLGATE: Solar Hybrid Gas Turbine Electric Power System,” Final Report, 2002.
[350] Behar, O., Khellaf, A., Mohammedi, K., Ait-Kaci, S., and Beghdadi, F., “Global CSP market
trends: the case of solar tower technology,” CIGE 2013, 2013.
[351] Bhargava, R. and Meher-Homji, C. B., “Parametric analysis of existing gas turbines with
inlet evaporative and overspray fogging,” ASME Turbo Expo 2002: Power for Land, Sea, and
Air, pp. 387-401), 2002.
[352] Thombare, D. G. and Verma, S. K., “Technological development in the Stirling cycle
engines,” Renewable and Sustainable Energy Reviews, 12, No. 1, pp. 1-38, 2008.
[353] Thombare, D. G. and Verma, S. K., “Technological development in the Stirling cycle
engines,” Renewable and Sustainable Energy Reviews, 12(1), 1-38, 2008.
[354]Urielie, I., “Ideal adiabatic analysis,” Ch. 4a, Lecture Notes – Ohio Univ.,
https://www.ohio.edu/mechanical/stirling/adiabatic/adiabatic.html
[355] Kalina, A.I., “Combined-cycle system with novel bottoming cycle,” J. Eng. Gas Turbines
Power, 106, pp. 737–42, 1984.
[356] Goswami, D.Y., Vijayaraghavan, S., Lu, S., Tamm, G., “New and emerging developments
in solar energy,” SolEnergy, 76, pp. 33–43, 2004.
[357] Kribus, A., “A high-efficiency triple cycle for solar power generation,” SolEnergy, 72, pp.
1–11, 2002.
[358] Ho, T., Mao, S.S. and Greif, R., “Comparison of the Organic Flash Cycle (OFC) to other
Advanced vapor cycles for intermediate and high temperature waste heat reclamation and
solar thermal energy,” Energy, 42, pp. 213–23, 2012.
[359] Shah, R.K., “Classification of heat exchangers” In Heat Echanger: Thermal-Hydraulic
Fundamentals and Design, S., Kakaç, Bergles, A.E. and Mayinger, F., Editors, Hemisphere
Publishing, Washington, D.C., pp. 9-46, 1981.
[360] Kakac, S., Liu, H. and Pramuanjaroenkij, A., “Gasketed-plate heat exchangers,” Ch. 11,
Heat exchangers: selection, rating, and thermal design, CRC press, 2012.
[361] Alfa Laval, “Gasketed plate-and-frame heat exchangers,” Accessed Aug. 2016,
http://www.alfalaval.us/products/heat-transfer/plate-heat-exchangers/Gasketed-plate-and-
frame-heat-exchangers/
[362] Vlasogiannis, P., Karagiannis, G., Argyropoulos, P., and Bontozoglou, V., “Air–water two-
phase flow and heat transfer in a plate heat exchanger,” Int. J. of Multiphase Flow, 28, No. 5,
pp. 757-772, 2002.

179
[363] Zohuri, B., “Application of Compact Heat Exchangers for Combined Cycle Driven
Efficiency in Next Generation Nuclear Power Plants: A Novel Approach,” Springer,
Switzerland, 2015.
[364] Kakac, S., Liu, H. and Pramuanjaroenkij, A., “Gasketed-plate heat exchangers,” Ch. 9, Heat
exchangers: selection, rating, and thermal design, CRC press, 2012.
[365] Shah, R.K. and Sekulic, “Classification of Heat Exchangers,” Fundamentals of Heat
Exchanger Design, Ch. 1, John Wiley & Sons, Hoboken, New Jersey, USA, 2003.
[366] Serna, M., and Jiménez, A., “A compact formulation of the Bell–Delaware method for heat
exchanger design and optimization,” Chemical Engineering Research and Design, 83, No. 5,
pp. 539-550, 2005.
[367] Selbaş, R., Kızılkan, Ö., and Reppich, M., “A new design approach for shell-and-tube heat
exchangers using genetic algorithms from economic point of view,” Chemical Engineering
and Processing: Process Intensification, 45, No. 4, pp. 268-275, 2006.
[368] Hesselgreaves, J. E., “Compact heat exchangers: selection, design and operation,” Gulf
Professional Publishing, Oxford, UK, 2001.
[369] Li, Q., Flamant, G., Yuan, X., Neveu, P., and Luo, L., “Compact heat exchangers: A review
and future applications for a new generation of high temperature solar receivers,” Renewable
and Sustainable Energy Reviews, 15, No. 9, pp. 4855-4875, 2011.
[370] Bengtson, H., “Features and characteristics of the flat plate heat exchanger,” Bright Hub
Engineering, Accessed Aug 2016, http://www.brighthubengineering.com/hvac/61791-
features-and-characteristics-of-the-flat-plate-heat-exchanger/#imgn_1
[371] Carlson, M. D., Kruizenga, A. K., Schalansky, C., and Fleming, D. F., “Sandia Progress on
Advanced Heat Exchangers for S-CO2 Brayton Cycles,” 4th Int. Symp. Supercritical CO2,
2, pp. 9-10, 2014.
[372] Wang, L., Sunden, B., Manglik, R.M., Plate heat exchangers: design, applications and
performance,” WIT Press, Southampton, UK, pp. 8-45, 2007.
[373] Lozano A, Barreras F, Fueyo N and Santodomingo S. “The flow in an oil/water plate heat
exchanger for the automotive industry,” Appl. Therm. Eng., 28, No. 10, pp. 1109–1117, 2008.
[374] Katzel, J., “Heat exchanger basics - A Plant Engineering Exclusive,” Plant Engineering,
2000.
[375] Durmus A, Benli H, Kurtbas I and Gul H. “Investigation of heat transfer and pressure drop
in plate heat exchangers having different surface profiles,” Int. J. Heat Mass Trans., 52, pp.
1451–7, 2009.
[376] Vermeulen, T. N. “Knowledge Sharing Report-CO 2 Liquid Logistics Shipping Concept
(LLSC),” Overall Supply Chain Optimization, Global CCS Institute, 2011
http://hub.globalccsinstitute.com/publications/co2-liquid-logistics-shipping-concept-llsc-
overall-supply-chain-optimization/53-co2
[377] Thonon B and Breuil E, “Compact heat exchangers technologies for HTRs recuperator
application. In: Proceedings of the technical committee meeting on gas turbine power
conversion systems for modular HTGRs,” IAEA-TECDOC, Palo Alto, CA, USA, pp. 1–11,
2001.
[378] Sherman, S.R., Chen, Y., “Heat exchanger testing requirements and facility needs for the
NHI/NGNP project,” WSRC-STI-2008-00152, STI, 2008.
[379] Zhang, L.W., Balachandar, S., Tafti, D.K. and Najjar, F.M., “Heat transfer enhancement
mechanisms in inline and staggered parallel-plate fin heat exchangers,” Int. J. Heat Mass
Trans., 40, No. 10, pp. 2307–25, 1997.

180
[380] Jacobi, A.M. and Shah, R.K., “Heat transfer surface enhancements through the use of
longitudinal vortices: a review of recent progress,” Exp. Therm. Fluid Sci., 11 pp. 295–309,
1995.
[381] Valencia, A., Fiebig, M. and Mitra, N.K., “Heat transfer enhancement by longitudinal
vortices in a fin-tube heat exchanger element with flat tubes,” J. Heat Trans., 118, pp. 209–
11, 1996.
[382] Mittal, R. and Balachandar, S., “Effect of three-dimensionality on the lift and drag of
nominally two-dimensional cylinders,” Phys. Fluids, 7, pp. 1841–65, 1995.
[383] Huang, L.J. and Shah, R.K., “Assessment of calculation methods for efficiency of straight
fins of rectangular profiles,” Int. J. Heat Fluid Flow, 13, pp. 282–293, 1992.
[384] Focke, W.W., Zachariades, J. and Oliver, I., “The effect of the corrugation inclination angle
on the thermo hydraulic performance of plate heat exchangers,” Int. J. Heat Mass Trans, 28,
No. 8, pp. 1469–79, 1985.
[385] Heatric, http://www.heatric.com/diffusion_bonded_exchangers.html
[386] Urquiza, E., Lee, K., Peterson, P. F., and Greif, R., “Multiscale transient thermal, hydraulic,
and mechanical analysis methodology of a printed circuit heat exchanger using an effective
porous media approach,” J. of Therm. Sci. & Eng. Apps., 5, No. 4, 2013.
[387] Pieve, M.A., “Calculation method for the optimized thermal-fluid dynamic sizing of heat
exchangers,” Heat Trans. Eng., 29, No. 6, pp. 556–64, 2008.
[388] Seo, J. W., Kim, Y. H., Kim, D., Choi, Y. D. and Lee, K. J., “Heat transfer and pressure drop
characteristics in straight microchannel of printed circuit heat exchangers,” Entropy, 17, No.
5, pp. 3438-3457, 2015.
[389] Dewson, S.J., “The development of high efficiency heat exchangers for helium gas cooled
reactors,” Paper No. 3213, Int. Conf. Adv.Nuclear Power Plants (ICAP), Cordoba, Spain;
2003.
[390] Ravindran, P., Sabharwall, P. and Anderson, N. A., “Modeling a printed circuit heat
exchanger with RELAP5-3D for the next generation nuclear plant,” Technical report, Idaho
National Laboratory, 2010.
[391] Sabharwall, P., “Engineering design elements of a two-phase thermosiphon to transfer
NGNP thermal energy to a hydrogen plant,” INL/EXT-09-15383, 2009.
[392] Li, X., Kininmont, D., Pierres, R.L. and Dewson, S.J., "Alloy 617 for the High Temperature
Diffusion- Bonded Compact Heat Exchangers," ICAPP 2008, pp. 7, 2008.
[393] Li, X., Le Pierres, R. and Dawson, S., "Heat exchangers for the next generation of nuclear
reactors," ICAPP 2006, 6, pp. 4–8, 2006.
[394] Kim, D.E., Kim, M.H., Cha, J.E. and Kim, S.O., “Numerical study for heat transfer and
pressure drop of supercritical carbon dioxide fluid with channel bending angle in Printed
Circuit Heat Exchanger,” Korean Society of Mechanical Engineering, Division of Thermal
Engineering; 2008.
[395] Gschwind, P., Regele, A., and Kottke, V. “Sinusoidal wavy channels with Taylor-Goertler
vortices,” Experimental thermal and fluid science, 11, No. 3, pp. 270-275, 1995.
[396] Sabharwall, P., Gunnerson, F., Tokuhiro A., Utgikar, V., Weaver, K. and Sherman S.,
“Theoretical design of a thermosyphon for a process heat removal from Next Generation
Nuclear Plant (NGNP) for production of hydrogen,” INL/EXT-07-13433, DOE, 2007.
[397] Mellouli, S., Askri, F., Dhaou, H., Jemni, A. and Ben Nasrallah, S. “A novel design of a heat
exchanger for a metal-hydrogen reactor,” Int. J. Hydrogen Energy, 32, No. 15, International
Symposium on Solar-Hydrogen-Fuel Cells, pp. 3501–3507, 2007.

181
[398] Picon-Nunez, M., Canizalez-Davalos, L., Martinez-Rodriguez, G. and Polley, G.T.,
“Shortcut design approach for spiral heat exchangers,” Food Bioproducts Process, 85, No. 4,
pp. 322–327, 2007.
[399] ASTM Standard C242-01, “Standard Terminology of Ceramic Whitewares and Related
Products,” ASTM International, West Conshohocken, PA, USA, 2007.
[400] Westabe, J., “Heat transfer international turns manure into power,” MLive Media Group,
2009, http://www.mlive.com/business/west-
michigan/index.ssf/2009/02/heat_transfer_international_tu.html
[401] Sommers, A., Wang, Q., Han, X., T’Joen, C., Park, Y. and Jacobi, A., “Ceramics and ceramic
matrix composites for heat exchangers in advanced thermal systems—a review,” Appl.
Therm. Eng. 30, pp. 1277–91, 2010.
[402] Lippmann, W., Knorr, J., Wolf, R., Rasper, R., Exner, H., Reinecke, A.-M., “Laser joining
of silicon carbide—a new technology for ultra-high temperature resistant joints,” Nucl. Eng.
Des., 231, pp. 151–161, 2004.
[403] Snead LL, Katoh Y, Kohyama A, Bailey JL, Vaughn NL, Lowden RA. Evaluation of neutron
irradiated near-stoichiometric silicon carbide fiber composites, Journal of Nuclear Materials,
Volumes 283–7, Part 1, 9th Int. Conf. on Fusion Reactor Materials, December; 2000. p. 551–
5.
[404] Greil, P., Lifka, T., and Kaindl, A., “Biomorphic cellular silicon carbide ceramics from
wood: II. Mechanical properties,” J. Eur. Ceram. Soc., 18, No. 14, pp.1975–1983, 1998.
[405] Steen M, Ranzani L. Potential of SiC as a heat exchanger material in combined cycle plant.
Ceram Int 2000;26(8):849–54.
[406] Amirthan G, Udaya kumar A, Balasubramanian M. Thermal conductivity studies on Si/SiC
ceramic composites. Ceram Int 2011;37(1):423–6.
[407] Alm, B., Imke, U., Knitter, R., Schygulla, U., and Zimmermann, S., “Testing and simulation
of ceramic micro heat exchangers,” Chem. Eng. J., 135, Supplement 1., pp. S179–S184, 2008.
[408] Islamoglu, Y., “Finite element model for thermal analysis of ceramic heat exchanger tube
under axial non-uniform convective heat transfer coefficient,” Mater Des., 25, No. 6, pp. 479–
82, 2004.
[409] Schulte-Fischedick, J., Dreißigacker, V., and Tamme, R., “An innovative ceramic high
temperature plate-fin heat exchanger for EFCC processes,” App. Therm. Eng., 27, No. 8, pp.
1285-1294, 2007.
[410] Islamoglu, Y., “Numerical analysis of the influence of a circular fin with different profiles
on the thermal characteristics in a ceramic tube of heat transfer equipment,” Int. J. Press.
Vessels Piping, 81, No. 7, pp. 583–587, 2004.
[411] Fend, T., Hoffschmidt, B., Pitz-Paal, R., Reutter, O. and Rietbrock, P., “Porous materials as
open volumetric solar receivers: experimental determination of thermophysical and heat
transfer properties,” Energy, 29, pp. 823–33, 2004.
[412] Fend, T., Pitz-Paal, R., Reutter, O., Bauer, J. and Hoffschmidt, B., “Two novel high porosity
materials as volumetric receivers for concentrated solar radiation,” Solar Energy Mater Solar
Cells, 84, pp. 291–304, 2004.
[413] Subramanian, R.S., “Flow through packed beds and fluidized beds,” Clarkson University,
2004, http://web2.clarkson.edu/projects/subramanian/ch301/notes/packfluidbed.pdf
[414] Szent-Gyorgyi, A., “Diffusion and Reaction,” Forler, Editor, Ch. 12, Elements of Chemical
Reaction Engineering, Univ. of Michigan, 2008.

182
[415] Suo, M., “Calculation methods for performance of heat exchangers enhanced with fluidized
beds,” Lett. Heat Mass Transfer, 3, pp. 555–564, 1976.
[416] Folyan, C.O., Pam, G.Y. and Habib, Y.B., “Fluidized bed combustion: Towards alternative
ways of energy recovery in rural Nigeria,” J. Energy Tech. & Policy, 5, No. 2, 2015.
[417] Vos, J., “Biomass Energy for Heating and Hot Water Supply in Belarus,” BTG Biomass
Technology Group BV, University of Twente, 2005.
[418] Gadgil P.R., “Survey on Fluidization Techniques and Its Advancements,” MGV’s Pharmacy
College, Nashik, 2009.
[419] Johari, A. and Taib, M.R., “Effect of Sand Sizes on the Fluidisation Behaviour in Circular
and Rectangular Columns,” J. of Chem. & Nat. Res. Eng., 2, pp. 14-21, 2007.
[420] Jeong, W. S., Lee, J. I., and Jeong, Y. H., “Potential improvements of supercritical
recompression CO 2 Brayton cycle by mixing other gases for power conversion system of a
SFR,” Nucl. Eng. & Design, 241, No. 6, pp. 2128-2137, 2011.
[421] Hexemer, M.J., “Supercritical CO2 Brayton Recompression Cycle Design and Control
Features to Support Startup and Operation,” sCO2 Power Cycle Symp., 2014.
[422] Edwards, M.F., Changal, A.A. and Parrott, D.L., “Heat transfer and pressure drop
characteristics of a plate heat exchanger using Newtonian and non-Newtonian liquids,” Chem.
Eng., pp. 286-293, 1974.
[423] Manglik, R.M., “Plate heat exchangers for process industry applications: enhanced
thermal-hydraulic characteristics of chevron plates,” Process Enhanced and Multiphase
Heat Transfer, Manglik, R.M. and Kraus, A.D., Eds., Begel House, New York, USA 1996.
[424] Muley, A. and Manglik, R.M., “Experimental investigation of heat transferenhancement in
a PHE with β = 60° chevron plates, in Heat and Mass Transfer, Tata McGraw-Hill, New
Delhi, India, 1995.
[425] Muley, A. and Manglik, R.M., “Enhanced heat transfer characteristics of singlephase flows
in a plate heat exchanger with mixed chevron plates,” J. Enhanced Heat Trans., 4, No. 3,
1997.
[426]. Muley, A., Manglik, R.M. and Metwally, H.M., “Enhanced heat transfer characteristicsof
viscous liquid flows in a chevron plate heat exchanger,” J. Heat Trans., 121, No. 4, 1999.
[427] Kakaç, S., Shah, R. K., and Aung, W., Handbook of Single-Phase Convective Heat
Transfer, John Wiley & Sons, New York, 1987.
[428] Fraas, A.P., “Heat exchanger design,” John Wiley & Sons, New Your, USA, 1989.
[429] Kern, D.Q., “Process heat transfer,” McGraw-Hill, New Yourk, USA, 1950.
[430] Petukhov, B.S., “Heat transfer and friction in turbulent pipe flow with variable physical
properties,” Advances in heat transfer, 6, pp. 503-564, 1970.
[431] Sieder, E.N. and Tate, G.E., “Heat transfer and pressure drop of liquids in tubes,”
Industrial and Engineering Chemistry, 28, pp. 1429-1435, 1936.
[432] Lang, C., and B. Lee, "Heat Transfer Fluid Life Time Analysis of Diphenyl
Oxide/Biphenyl Grades for Concentrated Solar Power Plants." Energy Procedia, 69, pp.
672-680, 2015.
[433] W.O. Walter, “Heat transfer technique with organic media,” In: Heat transfer media,
second ed. Graefelfing, Germany: Maria-Eich-Straβe; Ch.2, pp. 4–58, 1997.
[434] Ho, C.K., "Computational fluid dynamics for concentrating solar power systems," Wiley
Interdisciplinary Reviews: Energy and Environment, 3.3, pp.290-300, 2014.
[435] Cabeza, L.F., “Lithium in thermal energy storage: A state-of-the-art review,” Renewable
and Sustainable Energy Reviews, 42, pp. 1106-1112, 2015.

183
[436] Fernández, Á. G., Gomez, J. C., Galleguillos, H., & Fuentealba, E., “Thermophysical
properties and corrosion characterization of low cost lithium containing nitrate salts
produced in northern Chile for thermal energy storage,” SOLARPACES 2015, 1734, No. 1,
AIP Publishing, 2016.
[437] Mohanty, B. P. and Shores, D.A., “Role of Chlorides in Hot Corrosion of a Cast Fe–Cr–Ni
Alloy. Part I: Experimental Studies,” Corros. Sci., 46, No.12, pp. 2893–2907, 2004.
[438] Meier, G. H., and Pettit, F. S., 1989, “High-Temperature Corrosion of Alumina-Forming
Coatings for Superalloys,” Surf. Coat. Technol., 39/40, pp. 1–17.
[439] Ma, J., Jiang, S. M., Gong, J. and Sun, C., “Behavior and Mechanisms of Alkali-Sulfate-
Induced Hot Corrosion on Composite Coatings at 900°C,” Corros. Sci., 58, pp. 251–259,
2012.
[440] Glatzmaier, G. C. and Gomez, J.C., “Determining the Cost Benefit of High-Temperature
Coatings for Concentrating Solar Power Thermal Storage Using Probabilistic Cost
Analysis,” J. of Solar Energy Eng., 137, No. 4, 2015.
[441] Kruizenga, A. M., “Corrosion mechanisms in chloride and carbonate salts,” Sandia
National Laboratories, Livermore, CA Report No. SAND2012-7594, 2012.
[442] Pfleger, N., Bauer, T., Martin, C., Eck, M., and Wörner, A., “Thermal energy storage–
overview and specific insight into nitrate salts for sensible and latent heat storage,”
Beilstein Journal of Nanotechnology, 6, No. 1, pp. 1487-1497, 2015.
[443] Ho, C. K., Christian, J. M., Yellowhair, J., Ortega, J., & Andraka, C., “Fractal-like receiver
geometries and features for increased light trapping and thermal efficiency,”
SOLARPACES 2015, 1734, No. 1, AIP Publishing, 2016.
[444] Ortega, J. D., Khivsara, S. D., Christian, J. M., Yellowhair, J. E. and Ho, C. K., “Coupled
Optical-Thermal-Fluid Modeling of a Directly Heated Tubular Solar Receiver for
Supercritical CO2 Brayton Cycle,” ASME 2015 9th International Conference on Energy
Sustainability, American Society of Mechanical Engineers, San Diego, CA, USA, 2015.
[445] D. A. Kugath, G. Drenker, and A. A. Koenig, “Design and development of a paraboloidal
dish solar collector for intermediate temperature service,” Int. Solar Energy Soc., 1, pp.
449–453, Atlanta,Ga, USA, 1979.
[446] Nagarajan, P. K., Subramani, J., Suyambazhahan, S., & Sathyamurthy, R., “Nanofluids for
solar collector applications: a review,” Energy Procedia, 61, pp. 2416-2434, 2014.
[447] Frankel, A., Iaccarino, G., & Mani, A. “Convergence of the Bouguer–Beer law for
radiation extinction in particulate media,” J. Quant. Spec. & Radiative Trans., 182, pp.45-
54, 2016.
[448] Kostinksi, A., “On the extinction of radiation by a homogeneous but spatially correlated
random medium,” J. Opt.Soc. Am. A., 33, pp. 18-1929, 2001.
[449] Davis, A., Mineev-Weinstein, M., “Radiation propagation in random media: from positive
to negative correlations in high-frequency fluctuations,” J. Quant. Spectrosc. Radiation
Trans., 112, pp. 632–45, 2011.
[450] Coelho, P., Teerling, O. and Roekaerts, D., “Spectral radiative effects and
turbulence/radiation interaction in an on-luminous turbulent jet diffusion flame,” Combust
Flame, 133, pp. 75–91, 2003.
[451] Wu, Y., Haworth, D., Modest, M., Cuenot, B., “Direct numerical simulation of
turbulence/radiation interaction in premixed combustion systems,” Proc. Combustion Inst.,
30, pp. 639–46, 2005.

184
[452] Moens, L., and Blake, D. M., “Mechanism of hydrogen formation in solar parabolic trough
receivers,” J. Solar Energy Eng., 132, No. 3, 2010.
[453] de Miguel, M. T., Encinas-Sánchez, V., Lasanta, M. I., García-Martín, G. and Pérez, F. J.,
“Corrosion resistance of HR3C to a carbonate molten salt for energy storage applications in
CSP plants,” Solar Energy Materials and Solar Cells, 157, pp. 966-972, 2016.
[454] Ren, N., Wu, Y., Wang, T. and Ma, C. “Experimental study on optimized composition
of mixed carbonate for phase change thermal storage in solar thermal power plant,” J. Therm.
Anal. Calorim., 104, pp. 1201–1208, 2011.
[455] Olivares, R. Chen, C. and Wright, S., “The thermal stability of molten lithium-sodium-
potassium carbonate and the influence of additives on the melting point, J. Sol.Energy Eng.,
134, 2012.
[456] Flores, E., “A review of latent heat thermal energy storage for concentrated solar plants on
the grid,” J. Undergraduate Research, 5, No. 1, 2015.
[457] Stekli, J., Irwin, L. and Pitchumani, R., “Technical challenges and opportunities for
concentrating solar power with thermal energy storage,” J. Therm. Sci. & Eng. Apps., 5, No.
2, 2013.
[458] Bonan, G., “Ecological climatology,” Cambridge University Press, Cambridge U.K., 2002.

185
 10. APPENDIX
Appendix A.1. Plate-fin and tube-fin heat exchanger fin efficiencies for geometries of uniform fin
thickness [169].
 DISTRIBUTION

1 MS6123 Kenneth M. Armijo 6123

1 MS6123 Subhash Shinde 6123
1 MS0899 Technical Library 9536 (electronic copy)